# Open Source Models & Datasets

A technical concept covering open-source machine-learning models and the data information, code, and parameters needed to modify them.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Many Solutions, One Answer: Search Order Shapes the Default Solutions of Language Models

DevFeed: [Many Solutions, One Answer: Search Order Shapes the Default Solutions of Language Models](<https://devfeed.tech/articles/many-solutions-one-answer-search-order-shapes-the-default-solutions-of-language-models-145609.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25916>)

Author: Qinhan Hou, Jing Tang

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [Chain-of-thought](<https://devfeed.tech/topics/chain-of-thought.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [constraint](<https://devfeed.tech/tags/constraint.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [models](<https://devfeed.tech/tags/models.md>), [open-weight-models](<https://devfeed.tech/tags/open-weight-models.md>), [order](<https://devfeed.tech/tags/order.md>), [puzzle](<https://devfeed.tech/tags/puzzle.md>), [puzzles](<https://devfeed.tech/tags/puzzles.md>), [random](<https://devfeed.tech/tags/random.md>), [responses](<https://devfeed.tech/tags/responses.md>), [sampling](<https://devfeed.tech/tags/sampling.md>), [search](<https://devfeed.tech/tags/search.md>), [solutions](<https://devfeed.tech/tags/solutions.md>)

### AI overview

Across four open-weight language models, repeated answers to multi-solution puzzles cluster around a shared default solution far more than uniform sampling would predict. The models can still produce less common solutions when clues uniquely determine them. On held-out symbolic puzzles, a search visiting entities with the fewest initial candidates first better matched responses than name-order search when their predictions differed, by 26-40 percentage points. Changing value or search order shifted answers toward the favored solution. Randomized presentation or search order increased solution breadth by 26-42 points on symbolic puzzles and 11-26 on ZebraLogic, with maximum observed validity losses of 3 and 4 points. Randomized presentation broadened answers more than a random-choice instruction in each of the three main models.

### Source excerpt

arXiv:2611.25916v1 Announce Type: cross Abstract: When a constraint puzzle has several correct answers, language models tend to give the same one.We study four open-weight models from three families on symbolic puzzles and multi-solution ZebraLogic variants with exactly known solution sets.Repeated samples concentrate on one default solution far beyond uniform sampling, and different models often share it.Yet models usually produce even rarely sampled solutions when added clues uniquely identify them.On held-out symbolic puzzles, we compare visiting entities in name order with visiting those with the fewest initial candidates first. Both searches try values in the stated order. Where their pre-specified predictions differ, the latter matches a larger share of responses in all four models, by 26 to 40 percentage points.Reversing the stated value order or specifying a search order shifts responses towards the solution that the new order favours, on both testbeds. Randomising the presentation or supplying a random search order raises breadth (the effective share of solutions in use) by 26 to 42 percentage points on the symbolic task and 11 to 26 on ZebraLogic. The largest observed validity losses are 3 and 4 percentage points, respectively. Randomised presentation broadens answers more than a random-choice instruction in each of the three main models.

## Prior-Amortized In-Context Bayesian Inference for Generalized Linear Mixed-Effects Models

DevFeed: [Prior-Amortized In-Context Bayesian Inference for Generalized Linear Mixed-Effects Models](<https://devfeed.tech/articles/prior-amortized-in-context-bayesian-inference-for-generalized-linear-mixed-effects-models-144931.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2609.24422>)

Author: Alex Kipnis, Marcel Binz, Eric Schulz

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Predictive AI](<https://devfeed.tech/topics/predictive-ai.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [bayesian](<https://devfeed.tech/tags/bayesian.md>), [bayesian-inference](<https://devfeed.tech/tags/bayesian-inference.md>), [calibrated-uncertainty](<https://devfeed.tech/tags/calibrated-uncertainty.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [flow](<https://devfeed.tech/tags/flow.md>), [markov-chain-monte-carlo](<https://devfeed.tech/tags/markov-chain-monte-carlo.md>), [mcmc](<https://devfeed.tech/tags/mcmc.md>), [metropolis](<https://devfeed.tech/tags/metropolis.md>), [model-comparison](<https://devfeed.tech/tags/model-comparison.md>), [neural](<https://devfeed.tech/tags/neural.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [open-weights](<https://devfeed.tech/tags/open-weights.md>), [out-of-sample](<https://devfeed.tech/tags/out-of-sample.md>), [stat-ml](<https://devfeed.tech/tags/stat-ml.md>), [time](<https://devfeed.tech/tags/time.md>), [transformers](<https://devfeed.tech/tags/transformers.md>)

### AI overview

MetaBeta uses a pretrained amortized model for Bayesian inference in generalized linear mixed-effects models (GLMMs), taking the full prior and model specification as test-time context. It combines set transformers and conditional neural spline flows trained on simulated continuous, binary, and count data; Independence Metropolis-Hastings refinement against the unnormalized posterior is reported to yield tuning-free, asymptotically correct inference. As an alternative, MetaBeta warm-starts NUTS, with nearly identical inference reported 3.5-13 times faster than cold-started NUTS and over 98% fewer divergent transitions. The abstract reports agreement with NUTS on benchmarks and real datasets, including challenging misspecification and data-poor settings. A real-data prior-sensitivity analysis with model comparison reportedly took seconds with MetaBeta versus 14-18 core-hours with NUTS. These are claims from the supplied abstract; it does not provide methodological details or independent evaluation.

### Source excerpt

arXiv:2609.24422v3 Announce Type: replace Abstract: Hierarchical data is ubiquitous in the empirical sciences and most commonly analyzed with generalized linear mixed-effects models (GLMMs). Bayesian inference for GLMMs yields calibrated uncertainty estimates but requires Markov Chain Monte Carlo (MCMC). The gold-standard No-U-Turn Sampler (NUTS) is slow and must restart from scratch for every new dataset, model and prior specification. We introduce MetaBeta, a pretrained amortized model for in-context Bayesian inference over GLMMs. The full prior and model specification are supplied as context at test time, and the model returns an asymptotically exact posterior at the speed of a forward pass. Two set transformers and two conditional neural spline flows mirror the GLMM's hierarchical structure and are trained jointly on millions of realistic simulated datasets with continuous, binary and count outcomes. The flow posterior is refined by Independence Metropolis-Hastings against the unnormalized posterior. This yields tuning-free asymptotically correct inference. Alternatively, MetaBeta can warm-start NUTS. This gives nearly identical inference 3.5-13x faster than cold-started NUTS and removes over 98% of NUTS' divergent transitions. On controlled benchmarks with ground-truth parameters, MetaBeta matches NUTS in parameter recovery, posterior calibration and out-of-sample prediction. On real datasets with unknown parameters, priors and model structure, its posteriors closely agree with those of NUTS. Furthermore, it remains faithful where amortized models typically struggle: under misspecified likelihoods, misspecified priors, out-of-distribution predictors, collinear designs and data-poor regimes. On real data, a full prior sensitivity analysis including model comparison takes a few seconds with MetaBeta against 14-18 core-hours for NUTS, tracing its posteriors throughout. Our model is open-source and open-weights.

## COVERAGE-TRACE: Source-Side Counterfactual Tracing for Detecting Omissions in LLM Summaries

DevFeed: [COVERAGE-TRACE: Source-Side Counterfactual Tracing for Detecting Omissions in LLM Summaries](<https://devfeed.tech/articles/coverage-trace-source-side-counterfactual-tracing-for-detecting-omissions-in-llm-summaries-145534.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.25539>)

Author: Ali Karkehabadi, Jamshid Hassanpour, Rafael E. Banchs

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Factuality scoring in AI](<https://devfeed.tech/topics/factuality-scoring-in-ai.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [coverage](<https://devfeed.tech/tags/coverage.md>), [covering](<https://devfeed.tech/tags/covering.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [decision](<https://devfeed.tech/tags/decision.md>), [detection](<https://devfeed.tech/tags/detection.md>), [framework](<https://devfeed.tech/tags/framework.md>), [invariance](<https://devfeed.tech/tags/invariance.md>), [llm](<https://devfeed.tech/tags/llm.md>), [meaning](<https://devfeed.tech/tags/meaning.md>), [models](<https://devfeed.tech/tags/models.md>), [multiple](<https://devfeed.tech/tags/multiple.md>), [open-weight-models](<https://devfeed.tech/tags/open-weight-models.md>), [performance](<https://devfeed.tech/tags/performance.md>), [probability](<https://devfeed.tech/tags/probability.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [source](<https://devfeed.tech/tags/source.md>), [summary](<https://devfeed.tech/tags/summary.md>), [target](<https://devfeed.tech/tags/target.md>), [trace](<https://devfeed.tech/tags/trace.md>), [tracing](<https://devfeed.tech/tags/tracing.md>), [types](<https://devfeed.tech/tags/types.md>)

### AI overview

COVERAGE-TRACE detects potential omissions by replacing a designated atomic fact in the source with type-matched alternatives, keeping the candidate summary fixed, and measuring changes in its teacher-forced, length-normalized likelihood. Across 80 matched covered-omitted QMSum pairs and four open-weight models, its multi-counterfactual version reached 0.764-0.825 AUROC, above the strongest evaluated text-only baseline at 0.640. On Falcon-7B, targeted interventions scored 0.779 AUROC; shuffling counterfactual assignments lowered this to 0.533, and deleting the target helped while deleting unrelated material did not. Adjustment for meaning-preserving paraphrases reduced AUROC to 0.591-0.624. The method offers a complementary signal of source-summary dependence, but the authors caution it is not a wording-invariant probability of semantic coverage.

### Source excerpt

arXiv:2611.25539v1 Announce Type: new Abstract: Omissions are difficult to verify because missing information leaves no output span to inspect: a summary may be factually supported while still excluding a decision, qualification, quantity, or condition needed to preserve the source's meaning. We introduce COVERAGE-TRACE, a source-side counterfactual framework that evaluates omission as dependence on a designated atomic source fact. It replaces only the target fact with type-matched alternatives, holds the candidate summary fixed, and measures the resulting change in its teacher-forced, length-normalized likelihood. A summary covering the target should respond more strongly than one omitting it, while aggregation across multiple counterfactuals reduces sensitivity to replacement choice. We evaluate COVERAGE-TRACE on 80 matched covered--omitted QMSum pairs spanning several fact types and four open-weight models. Multi-CF achieves 0.764--0.825 AUROC, compared with 0.640 for the strongest evaluated text-only baseline. On Falcon-7B, targeted interventions achieve 0.779 AUROC, whereas shuffling counterfactual assignments reduces performance to 0.533; target deletion is effective while unrelated deletion is not. However, adjustment for meaning-preserving paraphrases reduces AUROC to 0.591--0.624. COVERAGE-TRACE therefore provides a practical, complementary signal of target-specific source--summary dependence, but not a wording-invariant semantic coverage probability. These results establish source-side intervention as a promising approach to omission detection while identifying semantic invariance as the central remaining challenge.

## Show Your Work: Can Small Open Models Read SEC Filings with Evidence a Machine Can Check?

DevFeed: [Show Your Work: Can Small Open Models Read SEC Filings with Evidence a Machine Can Check?](<https://devfeed.tech/articles/show-your-work-can-small-open-models-read-sec-filings-with-evidence-a-machine-can-check-145768.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24990>)

Author: Krishna Nandula

Published: 2026-11-20T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>)

Tags: [automatic](<https://devfeed.tech/tags/automatic.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [code](<https://devfeed.tech/tags/code.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [models](<https://devfeed.tech/tags/models.md>), [open](<https://devfeed.tech/tags/open.md>), [protocol](<https://devfeed.tech/tags/protocol.md>), [questions](<https://devfeed.tech/tags/questions.md>), [qwen3-5](<https://devfeed.tech/tags/qwen3-5.md>), [reasoning](<https://devfeed.tech/tags/reasoning.md>), [retry](<https://devfeed.tech/tags/retry.md>), [rtx-4090](<https://devfeed.tech/tags/rtx-4090.md>), [sec](<https://devfeed.tech/tags/sec.md>), [sec-filings](<https://devfeed.tech/tags/sec-filings.md>), [small](<https://devfeed.tech/tags/small.md>)

### AI overview

Filing-Evidence tests whether small open models can extract SEC filing figures with a page reference and verbatim quote. Across 1,000 questions from 90 recent 10-K and 10-Q filings for 45 companies, ten 7-21B instruct models ran on one RTX 4090; inline XBRL tags supplied the answer keys. Plain-setup accuracy with checkable evidence ranged from 17.1% to 82.7%, with Qwen3.5-9B reaching 84.5% in the best result. The checker rejected many incorrect answers, but 16-167 per 1,000 answerable questions still had a wrong value that passed every check. Among these plain-setup errors, 57% involved scale or units and 30% used the wrong period or column. Reasoning reduced these errors for four of five models offering it; retrying after a failed check improved accuracy by up to 3.9 points but did not reduce them. Questions, outputs, code, and protocol are public.

### Source excerpt

arXiv:2611.24990v1 Announce Type: new Abstract: A fluent answer from a language model is not a verified one. We ask whether small open models that run on a single consumer GPU can extract numbers from SEC filings with evidence a machine can check: the page and a verbatim quote containing the number. We introduce Filing-Evidence, a benchmark of 1,000 questions over 90 recent 10-K and 10-Q filings of 45 US companies, with answer keys taken from the filers' inline XBRL tags. We evaluate ten instruct models from ten developers (7-21B parameters) on one RTX 4090 in three setups: plain, check-and-retry, and the model's own reasoning mode. Accuracy with checkable evidence ranges from 17.1% to 82.7% in the plain setup, and the best result is 84.5% (Qwen3.5-9B). An automatic checker rejects many wrong answers, but between 16 and 167 answers per 1,000 give a wrong value for an answerable question and still pass every check. In the plain setup, 57% of these are scale or unit errors and 30% take another period or column of the right line. Reasoning significantly reduces such errors for 4 of 5 models that offer it. Retrying after a failed check raises accuracy by up to 3.9 points but never reduces them. Questions, model outputs, code and protocol are public.

## The Governance Gap in Open-Weight AI and What to Do About It

DevFeed: [The Governance Gap in Open-Weight AI and What to Do About It](<https://devfeed.tech/articles/the-governance-gap-in-open-weight-ai-and-what-to-do-about-it-143191.md>)

Original publisher: [Read original article](<https://www.securityweek.com/the-governance-gap-in-open-weight-ai-and-what-to-do-about-it/>)

Author: Etay Maor

Published: 2026-11-19T12:00:00Z

Content type: article

Language: en

Sources: [SecurityWeek](<https://devfeed.tech/sources/securityweek.md>)

Topics: [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [digital sovereignty](<https://devfeed.tech/topics/digital-sovereignty.md>), [software composition analysis](<https://devfeed.tech/topics/software-composition-analysis.md>)

Tags: [access-controls](<https://devfeed.tech/tags/access-controls.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-adoption](<https://devfeed.tech/tags/ai-adoption.md>), [ai-model](<https://devfeed.tech/tags/ai-model.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [attack](<https://devfeed.tech/tags/attack.md>), [attack-surface](<https://devfeed.tech/tags/attack-surface.md>), [automatically](<https://devfeed.tech/tags/automatically.md>), [backdoor](<https://devfeed.tech/tags/backdoor.md>), [backup](<https://devfeed.tech/tags/backup.md>), [data-privacy](<https://devfeed.tech/tags/data-privacy.md>), [digital-signature](<https://devfeed.tech/tags/digital-signature.md>), [fingerprint](<https://devfeed.tech/tags/fingerprint.md>), [governance](<https://devfeed.tech/tags/governance.md>), [gpus](<https://devfeed.tech/tags/gpus.md>), [hash](<https://devfeed.tech/tags/hash.md>), [hidden](<https://devfeed.tech/tags/hidden.md>), [it](<https://devfeed.tech/tags/it.md>), [open-weight-ai](<https://devfeed.tech/tags/open-weight-ai.md>), [risks](<https://devfeed.tech/tags/risks.md>)

### AI overview

Open-weight models can keep data within an organization's systems and give it more control over access and provider changes, but that control comes with responsibility for infrastructure, staffing, updates, and security. Risks described include malicious or altered model files, hidden behavior, delayed patching, and a larger attack surface; open weights also do not necessarily include training code or data details. The article recommends assigning an owner to each model, tracking its source, version, license, fine-tunes, and file hash, and testing models before deployment and upgrades. It also advises using a trusted internal repository, limiting model access, isolating higher-risk deployments, monitoring activity while protecting sensitive logs, and keeping a rollback version available.

### Source excerpt

Lower costs, data privacy, and greater control are driving open-weight AI adoption, but many organizations are not addressing the risks that come with it. The post The Governance Gap in Open-Weight AI and What to Do About It appeared first on SecurityWeek.

## Erased but Recoverable: On the Fragility of Concept Erasure Under Fine-Tuning

DevFeed: [Erased but Recoverable: On the Fragility of Concept Erasure Under Fine-Tuning](<https://devfeed.tech/articles/erased-but-recoverable-on-the-fragility-of-concept-erasure-under-fine-tuning-143146.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.24220>)

Author: Jun Li, Lizhi Xiong, Ziqiang Li, Yu Liu, Zenglin Shi

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [stable-diffusion](<https://devfeed.tech/topics/stable-diffusion.md>), [lora](<https://devfeed.tech/topics/lora.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [attention](<https://devfeed.tech/tags/attention.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [diffusion-models](<https://devfeed.tech/tags/diffusion-models.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [gradient](<https://devfeed.tech/tags/gradient.md>), [lora](<https://devfeed.tech/tags/lora.md>), [persistence](<https://devfeed.tech/tags/persistence.md>), [profiling](<https://devfeed.tech/tags/profiling.md>), [recovery](<https://devfeed.tech/tags/recovery.md>), [target](<https://devfeed.tech/tags/target.md>), [text-to-image](<https://devfeed.tech/tags/text-to-image.md>)

### AI overview

Concept erasure in text-to-image diffusion models can weaken or reverse after downstream fine-tuning such as DreamBooth, LoRA, or Custom Diffusion. The authors report that fine-tuning preserves certain high-contribution attention heads with stable output subspaces and response magnitudes, and that these heads in erased models can realign with their pretrained representations. Their proposed SHACE method profiles stability-critical heads and keeps their parameter blocks unchanged during erasure. Experiments reportedly show improved resistance to concept recovery across varied fine-tuning settings while maintaining target suppression and non-target generation quality.

### Source excerpt

arXiv:2611.24220v1 Announce Type: new Abstract: Existing concept erasure methods for text-to-image diffusion models primarily focus on suppressing target concepts while preserving non-target generation quality. However, in practical applications, open-source diffusion models are often further adapted through DreamBooth, LoRA, or Custom Diffusion. Such adaptations can reshape internal representations, thereby weakening or even reversing the erasure effects achieved by existing methods. We identify erasure persistence as a key yet underexplored dimension of concept erasure and analyze how downstream fine-tuning reshapes internal representations. Our analysis reveals that fine-tuning of the original model does not uniformly modify the model, but preserves certain high-contribution structures, particularly attention heads with stable output subspaces and response magnitudes. We also observe that, after downstream fine-tuning of concept-erased models, these heads' output representations more readily realign with their pretrained counterparts. Motivated by this observation, we propose Stable-Head-Aware Concept Erasure (SHACE), which improves erasure persistence by preserving stable attention structures during concept erasure. SHACE identifies stability-critical attention heads through Attention-Head Stability Profiling and keeps their corresponding parameter blocks unchanged during erasure via Stable-Head Gradient Preservation. Extensive experiments show that SHACE improves resistance to concept recovery across diverse fine-tuning settings while maintaining strong suppression and non-target generation quality.

## GridForge: An Open Physics-Informed Digital Twin and Benchmark Suite for UK 11 kV Underground Distribution Cables

DevFeed: [GridForge: An Open Physics-Informed Digital Twin and Benchmark Suite for UK 11 kV Underground Distribution Cables](<https://devfeed.tech/articles/gridforge-an-open-physics-informed-digital-twin-and-benchmark-suite-for-uk-11-kv-underground-distribution-cables-143366.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23241>)

Author: Lilliane Linnet Musoke

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Dataset development](<https://devfeed.tech/topics/dataset-development.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [catboost](<https://devfeed.tech/topics/catboost.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>), [Digital Public Good](<https://devfeed.tech/topics/digital-public-goods.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmark-dataset](<https://devfeed.tech/tags/benchmark-dataset.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [cables](<https://devfeed.tech/tags/cables.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [diagnosis](<https://devfeed.tech/tags/diagnosis.md>), [digital-twin](<https://devfeed.tech/tags/digital-twin.md>), [distribution](<https://devfeed.tech/tags/distribution.md>), [gradient-boosted-trees](<https://devfeed.tech/tags/gradient-boosted-trees.md>), [maintenance](<https://devfeed.tech/tags/maintenance.md>), [model](<https://devfeed.tech/tags/model.md>), [models](<https://devfeed.tech/tags/models.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [open](<https://devfeed.tech/tags/open.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [operational](<https://devfeed.tech/tags/operational.md>), [physics-informed](<https://devfeed.tech/tags/physics-informed.md>), [public](<https://devfeed.tech/tags/public.md>), [standards](<https://devfeed.tech/tags/standards.md>), [uk](<https://devfeed.tech/tags/uk.md>)

### AI overview

GridForge provides an open-source benchmark for fault diagnosis and prediction in UK 11 kV underground cables. It combines IEC 60287/60853 thermal models, a calibrated synthetic dataset covering 64 cables and 320 cable-years, and a PINN surrogate that enforces IEC 60287 algebraic balance. The PINN reports 0.148 °C RMSE against the analytical reference on a held-out set and trained in two minutes on the stated platform. Five benchmark tasks include pure-physics, gradient-boosted-tree, and PINN baselines, but only the virtual-sensor task distinguishes them on the reported test split; other tasks tie or are undefined. The dataset, test sets, models, and MIT-licensed code are made available for validation.

### Source excerpt

arXiv:2611.23241v1 Announce Type: new Abstract: The UK distribution-cable monitoring sector is data-rich but framework-poor: each Distribution Network Operator (DNO) collects megabytes of telemetry per asset per day, but there is no public benchmark dataset, no shared reference model, and no commonly agreed evaluation protocol for benchmarking sectoral, data-intensive models for operational decision support; notably for maintenance planning and prioritisation. We address this gap with GridForge, an open-source physics-informed digital twin and benchmark suite for fault diagnosis and prediction, integrating four building blocks: (i) a verified steady state and transient thermal model based on IEC 60287 and IEC 60853; (ii) a calibrated synthetic dataset of 64 cables (covering 320 cable-years) representing four canonical UK 11 kV thermal archetypes and four condition modes; (iii) a Physics-Informed Neural Network (PINN) surrogate that satisfies the algebraic balance defined by IEC 60287 during training and reaches an RMSE of 0.148 {\deg}C on an independent held out set against the reference analytical oracle; and (iv) a five-task benchmark suite that includes three reference baseline models (pure physics, gradient-boosted trees, and the PINN). All numerical parameters are sourced from public standards, the code is licensed under the MIT licence, and the trained PINN converges in only two minutes (as run on the computing platform in Section 10.1), enabling independent verification by DNO engineering and research practitioners. At present only the virtual-sensor task (T4) discriminates between the reference baselines; the other tasks tie or are undefined on this test split, so extending the baselines with task-specific logic is a priority future enhancement. For open validation, the dataset, test sets and models are made available at https://github.com/mlily2024/Gridforge

## A Rank-One Bypass of Fixed Linear Erasure

DevFeed: [A Rank-One Bypass of Fixed Linear Erasure](<https://devfeed.tech/articles/a-rank-one-bypass-of-fixed-linear-erasure-143460.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23880>)

Author: Jai Kumar Sharma, Vanshaj Khattar

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [cifar-10](<https://devfeed.tech/topics/cifar-10.md>)

Tags: [attack](<https://devfeed.tech/tags/attack.md>), [bypass](<https://devfeed.tech/tags/bypass.md>), [cifar-10](<https://devfeed.tech/tags/cifar-10.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [feature](<https://devfeed.tech/tags/feature.md>), [features](<https://devfeed.tech/tags/features.md>), [layer](<https://devfeed.tech/tags/layer.md>), [layers](<https://devfeed.tech/tags/layers.md>), [learning](<https://devfeed.tech/tags/learning.md>), [linear](<https://devfeed.tech/tags/linear.md>), [model](<https://devfeed.tech/tags/model.md>), [open](<https://devfeed.tech/tags/open.md>), [pair](<https://devfeed.tech/tags/pair.md>), [performance](<https://devfeed.tech/tags/performance.md>), [rank](<https://devfeed.tech/tags/rank.md>), [required](<https://devfeed.tech/tags/required.md>), [resistance](<https://devfeed.tech/tags/resistance.md>), [score](<https://devfeed.tech/tags/score.md>), [target](<https://devfeed.tech/tags/target.md>), [tasks](<https://devfeed.tech/tags/tasks.md>), [test](<https://devfeed.tech/tags/test.md>)

### AI overview

The paper shows that a fixed affine erasure layer can be bypassed if it transmits any direction visible to an allowed downstream linear classifier: an attacker can encode a target score in that direction using a rank-one affine update to the upstream model. Across six CIFAR-10 class-pair tasks, the method nearly restored pre-erasure target performance and could preserve a benign linear head's scores to numerical tolerance. In one constrained setting, eight target labels and 2,000 unlabelled calibration examples enabled 79.8% test accuracy with a 10% mean relative feature edit; tighter edit limits sharply reduced success. The authors frame linear guardedness as a guarantee about the current representation, not tamper resistance when open-model weights can be changed.

### Source excerpt

arXiv:2611.23880v1 Announce Type: new Abstract: Linear concept erasure can make a target unreadable to a specified family of linear classifiers applied to a model's current internal features. In an open-weight release, however, an attacker can modify the layers that produce those features. We study a fixed affine erasure layer whose upstream model is allowed to change. If the layer transmits even one direction visible to an allowed downstream classifier, the attacker can write a target score into that direction and carry it through the layer. We show that this condition is also necessary, and that for a nonconstant linear target score the required change is a rank-one affine update. Across six CIFAR-10 class-pair tasks, the route restores nearly all pre-erasure target performance and can leave a deployed benign linear head's scores unchanged to numerical tolerance. We then price the attack: with eight target labels and 2,000 unlabelled calibration examples, a 10% mean relative feature edit yields 79.8% test accuracy, which is 91.2% of the fitted score's gain over chance; tighter edit limits sharply reduce success. Linear guardedness therefore remains a valid guarantee about the current representation, but it is not by itself a tamper-resistance guarantee for an editable open-weight model.

## LusoClin: A Synthetic Clinical Note Dataset for Medical Information Retrieval in European Portuguese

DevFeed: [LusoClin: A Synthetic Clinical Note Dataset for Medical Information Retrieval in European Portuguese](<https://devfeed.tech/articles/lusoclin-a-synthetic-clinical-note-dataset-for-medical-information-retrieval-in-european-portuguese-142657.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23454>)

Author: Carla Teixeira Lopes, Daniel F\\'elix

Published: 2026-11-19T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [information retrieval](<https://devfeed.tech/topics/information-retrieval.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [lora](<https://devfeed.tech/topics/lora.md>), [llama](<https://devfeed.tech/topics/llama.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [managing large-scale text data in AI](<https://devfeed.tech/topics/managing-large-scale-text-data-in-ai.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Internationalization (i18n)](<https://devfeed.tech/topics/i18n.md>)

Tags: [bert](<https://devfeed.tech/tags/bert.md>), [clinical-search](<https://devfeed.tech/tags/clinical-search.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [document-retrieval](<https://devfeed.tech/tags/document-retrieval.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [information-retrieval](<https://devfeed.tech/tags/information-retrieval.md>), [languages](<https://devfeed.tech/tags/languages.md>), [llama](<https://devfeed.tech/tags/llama.md>), [low-resource-languages](<https://devfeed.tech/tags/low-resource-languages.md>), [medical](<https://devfeed.tech/tags/medical.md>), [note](<https://devfeed.tech/tags/note.md>), [resource](<https://devfeed.tech/tags/resource.md>), [synthetic](<https://devfeed.tech/tags/synthetic.md>)

### AI overview

LusoClin is a publicly released dataset of synthetic clinical notes in European Portuguese, intended to support privacy-preserving research on clinical information retrieval in a language with limited access to real clinical text. It was created by preprocessing real clinical data, fine-tuning Llama 3.2 (3B) with Quantized Low-Rank Adaptation, and generating notes from prompts; a BERT-based model automatically annotates clinical entities for search and indexing. Evaluations using automatic metrics and human review indicate that the dataset captures relevant clinical language and may support document retrieval and domain-specific model adaptation. However, annotators can often distinguish its notes from real ones, and some medical inconsistencies remain.

### Source excerpt

arXiv:2611.23454v1 Announce Type: new Abstract: Access to large-scale clinical text collections is crucial for advancing medical information retrieval research, including clinical note search, patient cohort identification, and domain-adapted retrieval models. However, real clinical notes are rarely available due to privacy regulations, which particularly limits progress for low-resource languages such as European Portuguese. We present LusoClin, a publicly released dataset of synthetic clinical notes in European Portuguese, created to support privacy-preserving research on clinical text retrieval. LusoClin was generated by pre-processing real clinical data, fine-tuning the open-source Llama 3.2 (3B) instruction-tuned model using Quantized Low-Rank Adaptation, and applying prompt-based clinical note generation. To facilitate semantic indexing and clinical search, the notes are automatically annotated with clinical entities using a BERT-based model. We evaluate LusoClin in terms of utility, fidelity, diversity, and privacy using automatic metrics and human evaluation. Results indicate that the dataset captures clinically relevant language patterns and can support downstream IR applications such as clinical document retrieval and domain-specific model adaptation, although annotators can often distinguish synthetic from real notes and some medical inconsistencies persist.

## An updated Clairic and Clairic-flash are now open-weight, from the new Workers AI Superintelligence team

DevFeed: [An updated Clairic and Clairic-flash are now open-weight, from the new Workers AI Superintelligence team](<https://devfeed.tech/articles/an-updated-clairic-and-clairic-flash-are-now-open-weight-from-the-new-workers-ai-superintelligence-team-142907.md>)

Original publisher: [Read original article](<https://blog.cloudflare.com/open-clairic/>)

Author: Kevin Flansburg

Published: 2026-11-18T22:00:00Z

Content type: release

Language: en

Sources: [Cloudflare Blog](<https://devfeed.tech/sources/cloudflare-blog.md>)

Topics: [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Cloudflare Workers](<https://devfeed.tech/topics/cloudflare-workers.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-superintelligence](<https://devfeed.tech/tags/ai-superintelligence.md>), [developers](<https://devfeed.tech/tags/developers.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [models](<https://devfeed.tech/tags/models.md>), [open](<https://devfeed.tech/tags/open.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [preview](<https://devfeed.tech/tags/preview.md>), [release](<https://devfeed.tech/tags/release.md>), [superintelligence](<https://devfeed.tech/tags/superintelligence.md>), [team](<https://devfeed.tech/tags/team.md>), [updated](<https://devfeed.tech/tags/updated.md>), [weight](<https://devfeed.tech/tags/weight.md>), [workers-ai](<https://devfeed.tech/tags/workers-ai.md>)

### AI overview

Cloudflare's updated Clairic and new Clairic-flash are open-weight extraction models hosted on Workers AI. Clairic is a redesigned 27B model; Clairic-flash is a compact 9B option. For ordinary text and number fields, Clairic uses parallel slot autoregressive generation, with fields advancing together and stopping independently; specialized heads handle estimates, forecasts, durations, calculated dates, and arithmetic. The redesign improves extraction from text, images, and PDFs while retaining confidence scores, character spans, and probabilities. Requests can cover up to four images or PDF pages, and line items can be returned as rows. Cloudflare reports sub-second HTTP responses and an internal image-reading evaluation improving from 207 errors to zero, but says its comparative extraction evaluations are self-reported and directional, with no public benchmarks available. Both models charge only for input tokens: Clairic-flash costs $0.10 per million input tokens, while Clairic is $0.24 after a promotion ending December 1, 2026; both are $0.10 until then.

### Source excerpt

Add your teaser here. It should complement the headline, not repeat it. Teaser cannot be longer than 240 characters. The Blog Review Gem will suggest an excerpt. Clairic graduates from preview, while Clairic-flash joins the repertoire on Workers AI. We're doing an open-weight release for both extraction models.

## App developers turn to QVAC for AI tooling without the cloud

DevFeed: [App developers turn to QVAC for AI tooling without the cloud](<https://devfeed.tech/articles/app-developers-turn-to-qvac-for-ai-tooling-without-the-cloud-141283.md>)

Original publisher: [Read original article](<https://www.networkworld.com/article/4250425/app-developers-turn-to-qvac-for-ai-tooling-without-the-cloud.html>)

Author: tether

Published: 2026-11-18T21:14:44Z

Content type: article

Language: en

Sources: [Network World](<https://devfeed.tech/sources/network-world.md>)

Topics: [Local AI](<https://devfeed.tech/topics/local-ai.md>), [On-device AI](<https://devfeed.tech/topics/on-device-ai.md>), [llama.cpp](<https://devfeed.tech/topics/llama-cpp.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [2026](<https://devfeed.tech/tags/2026.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-apps](<https://devfeed.tech/tags/ai-apps.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [api](<https://devfeed.tech/tags/api.md>), [app](<https://devfeed.tech/tags/app.md>), [apple-silicon](<https://devfeed.tech/tags/apple-silicon.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [artificial-intelligence-machine-learning](<https://devfeed.tech/tags/artificial-intelligence-machine-learning.md>), [backends](<https://devfeed.tech/tags/backends.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [costs](<https://devfeed.tech/tags/costs.md>), [cuda](<https://devfeed.tech/tags/cuda.md>), [data](<https://devfeed.tech/tags/data.md>), [desktop](<https://devfeed.tech/tags/desktop.md>), [developers](<https://devfeed.tech/tags/developers.md>), [development-kit](<https://devfeed.tech/tags/development-kit.md>), [device](<https://devfeed.tech/tags/device.md>), [edge](<https://devfeed.tech/tags/edge.md>), [edge-ai](<https://devfeed.tech/tags/edge-ai.md>), [edition-us](<https://devfeed.tech/tags/edition-us.md>), [engine](<https://devfeed.tech/tags/engine.md>), [language-en](<https://devfeed.tech/tags/language-en.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [storytype-brandpost](<https://devfeed.tech/tags/storytype-brandpost.md>), [tooling](<https://devfeed.tech/tags/tooling.md>)

### AI overview

Tether's QVAC SDK combines local model inference and fine tuning with tools for offline retrieval-augmented generation, aiming to reduce cloud dependence, per-request costs, and exposure of user data. Its Fabric engine builds on llama.cpp and GGML, with Vulkan, Apple Silicon's Metal and Core ML, and CUDA backends. The SDK's TurboVec integration is described as reducing embedding memory use by up to 8x with 4-bit quantization, helping vector search fit consumer hardware. QVAC also supports quantized open-source models and Tether's compact Psy models; the source cites an 800-million-parameter African-language translation model outperforming a 27-billion-parameter TranslateGemma on key benchmarks. These claims present local AI as more accessible, though the excerpt does not provide benchmark details or specify hardware requirements.

### Source excerpt

In most instances today, adding AI to an app means sending user data to a third-party API, paying per request, and hoping the connection holds. Tether's QVAC is built on the opposite premise. It runs AI models directly on the user's device. Data never leaves the hardware, there are no per-request costs, no rate limits, and no need for an internet connection once a model is downloaded. QVAC SDK: the next milestone in cloud-free AI While open-source tools and standards like llama.cpp, GGML, and ONNX have made local AI much more accessible in recent years, creating new possibilities for consumer devices, the edge AI ecosystem remains fragmented. Against this backdrop, in 2026, Tether released a software development kit (SDK) that pulls together the best open tooling from across different domains of machine learning. The centerpiece of Tether's QVAC SDK is Fabric. Built on llama.cpp, it expands the engine to enable on-device fine tuning and additional AI capabilities--opening access to memory-intensive workloads previously only possible with expensive hardware or cloud services. Building on GGML's modular framework, QVAC Fabric is primarily oriented toward three backends: Vulkan, which covers most consumer-grade desktop and mobile chips; Metal and Core ML for Apple Silicon; and CUDA for higher-end NVIDIA GPU setups. The QVAC SDK empowers developers with a unified interface for building multiplatform AI apps that run on-device. But although significant progress has been made in bringing inference to consumer devices, text generation alone is only one part of the pipeline that makes modern AI useful. Bringing search to edge AI RAG (retrieval-augmented generation) is what equips AI to import and process contextual information to enhance its output. Instead of language models relying solely on their training data to answer questions, RAG lets them answer using a specific set of documents. Implementing RAG into an AI app incurs significant memory overhead versus running a mod

## Two decades of The Apache Way: Celebrating 20 years of Google and the Apache Software Foundation

DevFeed: [Two decades of The Apache Way: Celebrating 20 years of Google and the Apache Software Foundation](<https://devfeed.tech/articles/two-decades-of-the-apache-way-celebrating-20-years-of-google-and-the-apache-software-foundation-144106.md>)

Original publisher: [Read original article](<http://opensource.googleblog.com/2026/11/two-decades-of-the-apache-way-google-asf-20-years.html>)

Author: Google Open Source (noreply@blogger.com)

Published: 2026-11-18T19:30:00Z

Content type: article

Language: en

Sources: [Google Open Source Blog](<https://devfeed.tech/sources/google-open-source-blog.md>)

Topics: [Open Source](<https://devfeed.tech/topics/open-source.md>), [open source maintenance](<https://devfeed.tech/topics/open-source-maintenance.md>), [engineering-culture](<https://devfeed.tech/topics/engineering-culture.md>), [Cloud Dataflow](<https://devfeed.tech/topics/cloud-dataflow.md>), [Apache Spark](<https://devfeed.tech/topics/spark.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [gemma](<https://devfeed.tech/topics/gemma.md>)

Tags: [2006](<https://devfeed.tech/tags/2006.md>), [2016](<https://devfeed.tech/tags/2016.md>), [ai](<https://devfeed.tech/tags/ai.md>), [airflow](<https://devfeed.tech/tags/airflow.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apache](<https://devfeed.tech/tags/apache.md>), [apache-2-0](<https://devfeed.tech/tags/apache-2-0.md>), [apache-airflow](<https://devfeed.tech/tags/apache-airflow.md>), [apache-software-foundation](<https://devfeed.tech/tags/apache-software-foundation.md>), [apache-spark](<https://devfeed.tech/tags/apache-spark.md>), [asf](<https://devfeed.tech/tags/asf.md>), [batch](<https://devfeed.tech/tags/batch.md>), [beam](<https://devfeed.tech/tags/beam.md>), [big-data](<https://devfeed.tech/tags/big-data.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [cloud-dataflow](<https://devfeed.tech/tags/cloud-dataflow.md>), [code](<https://devfeed.tech/tags/code.md>), [collaboration](<https://devfeed.tech/tags/collaboration.md>), [communities](<https://devfeed.tech/tags/communities.md>), [community](<https://devfeed.tech/tags/community.md>), [google](<https://devfeed.tech/tags/google.md>), [google-supports-oss](<https://devfeed.tech/tags/google-supports-oss.md>), [open-source](<https://devfeed.tech/tags/open-source.md>)

### AI overview

Google's relationship with the Apache Software Foundation (ASF), formalized in 2006, continues through sponsorship, upstream contributions, and shared support for open, vendor-neutral, community-led development. Google says its contributor license agreements draw on ASF models and Apache 2.0 is its preferred code license, also used for open hardware and models such as Gemma 4. Examples of collaboration include donating Dataflow SDK components that helped establish Apache Beam; managed Google Cloud services built on Airflow and Spark; and BigQuery support for Apache Iceberg. Google was the ASF's first Platinum Sponsor and contributes funding and cloud infrastructure; Googlers also serve as committers and PMC members, while ASF has participated in Google Summer of Code for over 20 years. The text points to open data, cloud infrastructure, and AI as areas for future collaboration.

### Source excerpt

by Mike Bufano, Open Source Programs Office In 2006, Google formalized its relationship with the Apache Software Foundation (ASF) as one of its first corporate sponsors and twenty years later, that partnership continues through ongoing sponsorship and upstream contributions. At the core of our collaboration is a shared belief in "The Apache Way" and its principles of open, vendor-neutral, and community-led development. From the early days of corporate open source to modern machine learning and open data systems, Google and ASF communities have collaborated on widely used open source software. Shaping how Google does open source When we established Google's open source contribution workflow, we drew on established ASF practices. Both our Individual and Corporate Contributor License Agreements (CLAs), which external developers sign before contributing to a Google-owned open source project, are based on the ASF's CLAs. The Apache 2.0 license also remains Google's preferred open source license for code. Alongside our traditional software releases, we also use Apache 2.0 for open hardware designs and open models like Gemma 4. Co-engineering upstream and in the cloud Across Google Cloud, our engineering teams contribute code upstream alongside Apache project communities and build managed services directly around those open standards: Google Cloud Dataflow & Apache Beam: Drawing on Google's internal Flume and Dataflow systems, Google donated the original Dataflow SDK and runners to the Apache Incubator in 2016 to help establish Apache Beam as a vendor-neutral open standard for unified batch and stream processing. Google Cloud Managed Service for Apache Airflow & Apache Airflow: Managed workflow orchestration built on Apache Airflow, helping engineering and data teams author, schedule, and monitor complex pipelines. Google Cloud Managed Service for Apache Spark & Apache Spark: Managed big data and analytics processing built on Apache Spark, running directly on Google's clou

## Wayfinder: Frontier-level Code Search at a Fraction of the Cost

DevFeed: [Wayfinder: Frontier-level Code Search at a Fraction of the Cost](<https://devfeed.tech/articles/wayfinder-frontier-level-code-search-at-a-fraction-of-the-cost-141039.md>)

Original publisher: [Read original article](<https://blog.voyageai.com/2026/11/18/wayfinder-code-search/>)

Author: Voyage AI

Published: 2026-11-18T15:30:00Z

Content type: article

Language: en

Sources: [Voyage AI](<https://devfeed.tech/sources/voyage-ai.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [hybrid-search](<https://devfeed.tech/topics/hybrid-search.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [hierarchical agent systems](<https://devfeed.tech/topics/hierarchical-agent-systems.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [code](<https://devfeed.tech/tags/code.md>), [code-search](<https://devfeed.tech/tags/code-search.md>), [cost](<https://devfeed.tech/tags/cost.md>), [frontier](<https://devfeed.tech/tags/frontier.md>), [grep](<https://devfeed.tech/tags/grep.md>), [news](<https://devfeed.tech/tags/news.md>), [search](<https://devfeed.tech/tags/search.md>), [wayfinder](<https://devfeed.tech/tags/wayfinder.md>)

### AI overview

Wayfinder pairs a small fine-tuned code-search model with a local hybrid index to make agent search cheaper while finding more relevant files. The index combines semantic search (80%) with BM25 (20%); its results are merged and reranked with grep results, then the model runs follow-up greps. On 200 held-out tasks, open models with this approach found about 74% of the right files for roughly $0.01 per search, compared with 76% for Sonnet 5.5 high effort at $0.12 and 78% for GPT-6.1 Sol medium effort at $0.07. Two pooled Wayfinder runs reached 78% for about $0.02. In a 42-task coding evaluation, competitive search methods had similar pass rates; Wayfinder reached 55%, while pooled runs reached 61%, but the sample was too small to establish a clear lead. Indexing also has ongoing costs: the authors estimate a typical repository pays back after about two searches monthly. Wayfinder is an open-source research project, not a MongoDB product; the authors recommend evaluating it on individual codebases.

### Source excerpt

Author: Henry Weller Before a coding agent can fix a bug, it has to find the code that matters. The coding agent industry has generally embraced grep as the way agents search code, but repeated searches with grep can become slow and expensive as codebases grow. For customers building coding agents, we wanted to explore [...]

## An Open-Source Training Dataset for Foundation Models for Hyperparameter Optimization

DevFeed: [An Open-Source Training Dataset for Foundation Models for Hyperparameter Optimization](<https://devfeed.tech/articles/an-open-source-training-dataset-for-foundation-models-for-hyperparameter-optimization-141568.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.23417>)

Author: Aaron Klein, Herilalaina Rakotoarison, Luca Thale-Bombien, David Salinas

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Hyperparameter optimization](<https://devfeed.tech/topics/hyperparameter-optimization.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [scaling laws](<https://devfeed.tech/topics/scaling-laws.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [hyperparameter-optimization](<https://devfeed.tech/tags/hyperparameter-optimization.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [training](<https://devfeed.tech/tags/training.md>), [training-dataset](<https://devfeed.tech/tags/training-dataset.md>)

### AI overview

BBO-Pile is an open-source dataset of over 500,000 optimization trajectories across 3,095 hyperparameter optimization tasks and global optimization problems, addressing the shortage of large, public pre-training data. The authors train foundation models ranging from 2M to 80M parameters and 200M to 2B training tokens, and study scaling with compute. They report that large-scale pre-training can effectively imitate hyperparameter optimization methods. The supplied text does not detail specific performance comparisons or limitations.

### Source excerpt

arXiv:2605.23417v2 Announce Type: replace Abstract: Most hyperparameter optimization methods require extensive tuning of their own hyperparameters, often limiting their ability to generalize across different benchmark families. Foundation models for hyperparameter optimization that learn optimization principles from a large collection of optimization trajectories offer a promising alternative, with the potential to outperform manually designed methods across diverse problem classes. However, prior work as either relied on non-public datasets or on purely synthetic data, limiting reproducibility and generalization to real-world problems. As a result, progress in this area has been constrained by the lack of large-scale, real-world, publicly available pre-training data. We introduce BBO-Pile, the first open-source dataset comprising over 500K optimization trajectories evaluated across 3,095 different hyperparameter optimization tasks and global optimization problems for different optimizers, which represents by far the largest public dataset for this task. Using this dataset, we train a family of foundation models at multiple scales, ranging from 2M to 80M parameters and from 200M to 2B training tokens, and study their scaling behavior with respect to compute. Our results demonstrate that large-scale pre-training is a viable and effective approach to imitate hyperparameter optimization methods, paving the way for future research in this direction.

## AutoE2E: Don't speak - Language free end-to-end autonomy A Real-Time Driving Model for L2++ ADAS and L4 Robotaxis

DevFeed: [AutoE2E: Don't speak - Language free end-to-end autonomy A Real-Time Driving Model for L2++ ADAS and L4 Robotaxis](<https://devfeed.tech/articles/autoe2e-don-t-speak-language-free-end-to-end-autonomy-a-real-time-driving-model-for-l2-adas-and-l4-robotaxis-140020.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22374>)

Author: Muhammad Zain Khawaja, Zaynab El Mawas, Ryota Yamada, Gabriela Cordova, Arseni Lysak, Intisar Chowdhury, Bharat Jayaprakash, Apostolos Georgiadis, Shri Harsha Adapala, Jatin Gautam, Christoph Bank, Al

Published: 2026-11-18T05:00:00Z

Content type: release

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [autonomous vehicles](<https://devfeed.tech/topics/autonomous-vehicles.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [NVIDIA DRIVE](<https://devfeed.tech/topics/nvidia-drive.md>)

Tags: [adas](<https://devfeed.tech/tags/adas.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-driving](<https://devfeed.tech/tags/autonomous-driving.md>), [autonomy](<https://devfeed.tech/tags/autonomy.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [camera](<https://devfeed.tech/tags/camera.md>), [checkpoint](<https://devfeed.tech/tags/checkpoint.md>), [code](<https://devfeed.tech/tags/code.md>), [collaboration](<https://devfeed.tech/tags/collaboration.md>), [community](<https://devfeed.tech/tags/community.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [dashboard](<https://devfeed.tech/tags/dashboard.md>), [driving](<https://devfeed.tech/tags/driving.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [experimentation](<https://devfeed.tech/tags/experimentation.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fde](<https://devfeed.tech/tags/fde.md>), [model](<https://devfeed.tech/tags/model.md>), [real-time](<https://devfeed.tech/tags/real-time.md>), [robotaxis](<https://devfeed.tech/tags/robotaxis.md>)

### AI overview

AutoE2E is an open-source end-to-end driving model that combines multi-camera perception, HD-map information and route conditioning in a temporal bird's-eye-view architecture to predict vehicle motion. Its v1.0 release includes checkpoints, training and evaluation code, and DataModelConsole for inspecting experiments. On KITScenes validation, the fine-tuned checkpoint reports 0.745 m average displacement error and 2.020 m final displacement error over a three-second horizon. The paper also describes preliminary real-vehicle research on TUM's EDGAR platform and ongoing collaboration with KIT. The release is an early research baseline: quantitative results are open-loop, and validated closed-loop autonomous driving has not been demonstrated.

### Source excerpt

arXiv:2611.22374v1 Announce Type: new Abstract: AutoE2E is an open-source end-to-end driving model developed by the Autoware Foundation Robotaxi Working Group. It introduces a temporal bird's-eye-view (BEV) architecture that combines multi-camera perception, HD-map information and route conditioning to predict future vehicle motion. The evaluated model uses a pretrained camera-BEV backbone, temporal feature fusion and a lightweight recurrent trajectory planner. It contains approximately 79.9 million parameters, of which approximately 1.45 million are trainable in the evaluated configuration. The v1.0 release provides model checkpoints, training and evaluation code, and DataModelConsole, a dashboard for visualizing and inspecting model experiments. On the KITScenes validation benchmark, the fine-tuned checkpoint achieves an average displacement error (ADE) of 0.745 m and a final displacement error (FDE) of 2.020 m at a three-second prediction horizon. The paper also reports preliminary real-vehicle research conducted with the Technical University of Munich's EDGAR platform, alongside ongoing collaboration with the Karlsruhe Institute of Technology. AutoE2E v1.0 is an early open research release. Its quantitative evaluation is open-loop, and the model has not yet demonstrated validated closed-loop autonomous driving. The release is intended to provide the autonomous-driving community with an inspectable baseline for further research, experimentation and collaborative development.

## Watermark Evidence under Open-Weight LLM Modification: Exact Bounds for Low-Rank Output Watermarks

DevFeed: [Watermark Evidence under Open-Weight LLM Modification: Exact Bounds for Low-Rank Output Watermarks](<https://devfeed.tech/articles/watermark-evidence-under-open-weight-llm-modification-exact-bounds-for-low-rank-output-watermarks-139900.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21702>)

Author: Prashant Shekhar, Dumindu Samaraweera, Caroline Howard

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Cryptography and Security](<https://devfeed.tech/sources/arxiv-cryptography-and-security.md>)

Topics: [Large Language Model](<https://devfeed.tech/topics/llm.md>), [lora](<https://devfeed.tech/topics/lora.md>), [quantization](<https://devfeed.tech/topics/quantization.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [attack](<https://devfeed.tech/tags/attack.md>), [covariance](<https://devfeed.tech/tags/covariance.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [false-positives](<https://devfeed.tech/tags/false-positives.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [int8](<https://devfeed.tech/tags/int8.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [llama](<https://devfeed.tech/tags/llama.md>), [lower-bound](<https://devfeed.tech/tags/lower-bound.md>), [mistral](<https://devfeed.tech/tags/mistral.md>), [robustness](<https://devfeed.tech/tags/robustness.md>), [sequence](<https://devfeed.tech/tags/sequence.md>), [span](<https://devfeed.tech/tags/span.md>), [theoretical](<https://devfeed.tech/tags/theoretical.md>), [watermark](<https://devfeed.tech/tags/watermark.md>)

### AI overview

The study derives bounds on how changes to open-weight LLM logits affect OpenStamp watermark evidence, including a residual-span lower bound, a covariance-based estimate of evidence drift, and sequence-level bounds on miss probability after attack selection. In tests on Llama-2-7B and Mistral-7B-v0.3, INT8 quantization and 30% pruning retained true-positive rates of 85.6-94.8%, while 500-step LoRA fine-tuning reduced rates to 28.8% and 11.6%. The covariance estimate tracked measured LoRA evidence drift, with no remainder-bound violations across 160,000 test prefixes. The experiments also recorded zero false positives among 500 base-model continuations per model.

### Source excerpt

arXiv:2611.21702v1 Announce Type: cross Abstract: Open-weight language models can continue generating after their released weights are modified, while a watermark verifier still scores outputs against the original base and marked models. We study how such modifications change the evidence available to an OpenStamp-style verifier. At each shared prefix, we represent the modification as a residual in the next-token logits. We derive an exact finite-vocabulary lower bound on the expected log-likelihood-ratio evidence under a residual-span constraint, along with a sharp span-only deviation bound. We also characterize the first-order evidence shift through a covariance term with an explicit second-order remainder, and determine the residual span needed to reach a target evidence level. For sequence-level detection, we derive an exact conditional exponential-moment bound and combine it with a supermartingale argument and a distribution-free order-statistic tolerance bound. This yields a high-confidence upper bound on miss probability after attack selection on calibration data and evaluation on independent certification continuations. In experiments with OpenStamp on Llama-2-7B and Mistral-7B-v0.3, INT8 quantization and 30% pruning retain test true-positive rates of 85.6-94.8%, while 500-step LoRA fine-tuning lowers them to 28.8% and 11.6%. The covariance prediction tracks the measured mean LoRA evidence drift on both models, with no remainder-bound violations across 160,000 test prefixes. Additionally, we observe zero false positives among 500 base-model continuations per model. Through these theoretical and empirical results, the current research provides a better understanding of watermark robustness to model-editing/tuning attacks.

## Long-Context Generation Is a Sampling Problem (When You Turn Up the Heat)

DevFeed: [Long-Context Generation Is a Sampling Problem (When You Turn Up the Heat)](<https://devfeed.tech/articles/long-context-generation-is-a-sampling-problem-when-you-turn-up-the-heat-139971.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.22148>)

Author: Allen G. Roush, Minh Nhat Nguyen, Ravid Shwartz-Ziv, Judah Goldfeder, Sanjay Basu

Published: 2026-11-18T05:00:00Z

Content type: news

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [Language models](<https://devfeed.tech/topics/language-models.md>), [Temperature settings in AI](<https://devfeed.tech/topics/temperature-settings-in-ai.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [context window](<https://devfeed.tech/topics/context-window.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [chain-of-thought](<https://devfeed.tech/tags/chain-of-thought.md>), [code](<https://devfeed.tech/tags/code.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [correlation](<https://devfeed.tech/tags/correlation.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [data](<https://devfeed.tech/tags/data.md>), [generation](<https://devfeed.tech/tags/generation.md>), [github](<https://devfeed.tech/tags/github.md>), [github-issues](<https://devfeed.tech/tags/github-issues.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [llama](<https://devfeed.tech/tags/llama.md>), [llama-3-1](<https://devfeed.tech/tags/llama-3-1.md>), [long-context](<https://devfeed.tech/tags/long-context.md>), [open-weight-models](<https://devfeed.tech/tags/open-weight-models.md>), [qwen3-5](<https://devfeed.tech/tags/qwen3-5.md>), [reading](<https://devfeed.tech/tags/reading.md>), [sampling](<https://devfeed.tech/tags/sampling.md>)

### AI overview

The study reports that sampler behavior as temperature rises can strongly affect long-generation quality in open-weight models. Across 29 models and ten benchmarks plus retrieval up to one million tokens, top-nsigma retained a median 100% of its temperature-1 score at temperature 2, while top-p retained 26%. On three Qwen3.5 models, coding agents using top-p, top-k, or Top-H resolved none of 100 GitHub issues at temperature 2; top-nsigma resolved 46 to 62. The authors associate quality loss with growth in the candidate-token pool: for one Llama-3.1-8B example, top-p expanded from 4 tokens at temperature 1 to 51,845 at 2, while top-nsigma stayed at 3. They report that fixing the pool at temperature 1 also prevented Top-H, min-p, and p-less from collapsing. Most samplers performed similarly at temperature 1.

### Source excerpt

arXiv:2611.22148v1 Announce Type: new Abstract: Open-weight language models are widely observed to get worse the longer they generate. Often, attempts to fix this involve more training. At high temperature, The sampler, i.e. the rule that picks the next token out of the model's probabilities, decides whether long and agentic generation survives and/or thrives. We test 29 open-weight models from 1B to 122B parameters on ten established benchmarks plus needle retrieval out to one million tokens, covering SWE-bench Verified coding agents, long chain-of-thought reasoning, long-form writing, long structured output and long-context reading. Samplers that keep their pool of candidate tokens small and steady as temperature rises don't lose accuracy. Samplers that let in low-probability tokens fall apart. At temperature 2, coding agents running top-p, top-k or Top-H resolve 0 of 100 GitHub issues, while the same agents running top-nsigma resolve 46 to 62 and win every issue the two samplers disagree on, on three Qwen3.5 models. Over all benchmarks top-nsigma keeps a median 100% of its temperature-1 score where top-p keeps 26%. Counting how many tokens each sampler lets through explains why. In one Llama-3.1-8B story, going from temperature 1 to 2 grows top-p's pool from 4 tokens to 51,845 while top-nsigma stays at 3. Across 133 combinations of sampler setting and temperature, the typical pool size predicts judged quality (Spearman correlation of -0.77), and when we pick the pool at temperature 1 and only apply temperature afterwards, Top-H, min-p and p-less stop collapsing. At the default temperature of 1 most samplers perform about the same, We release all code and data to the public.

## OpenMHC: Accelerating the Science of Wearable Foundation Models

DevFeed: [OpenMHC: Accelerating the Science of Wearable Foundation Models](<https://devfeed.tech/articles/openmhc-accelerating-the-science-of-wearable-foundation-models-139792.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2607.16235>)

Author: Narayan Schuetz, Yuze Bai, Lianggang Pan, Edgar Eggert, Favour Nerrise, Juan Delgado-SanMartin, Max Rosenblattl, Milana Gurbanova, Mohammad Asadi, Anders Johnson, Steven G. Hershman, Paul Schmiedmayer

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [devices](<https://devfeed.tech/tags/devices.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [health](<https://devfeed.tech/tags/health.md>), [health-ai](<https://devfeed.tech/tags/health-ai.md>), [health-monitoring](<https://devfeed.tech/tags/health-monitoring.md>), [model-weights](<https://devfeed.tech/tags/model-weights.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [wearable](<https://devfeed.tech/tags/wearable.md>), [wearable-health](<https://devfeed.tech/tags/wearable-health.md>), [xgboost](<https://devfeed.tech/tags/xgboost.md>)

### AI overview

OpenMyHeartCounts (OpenMHC) combines wearable data, open model implementations and weights, and a benchmark to support reproducible wearable health research. The dataset, derived from more than a decade of the My Heart Counts app, contains over 67 million participant-hours from 11,894 participants, with up to 19 sensor channels and 170 health, lifestyle, mood, and behavioral variables. The benchmark covers outcome prediction, multivariate imputation, and time-series forecasting. Foundation models perform strongly, but carefully tuned classical baselines such as XGBoost remain highly competitive. Access is described as broad for qualified researchers.

### Source excerpt

arXiv:2607.16235v4 Announce Type: replace-cross Abstract: Mobile and wearable devices offer an unprecedented opportunity for continuous, passive health monitoring and active health coaching. However, the largest wearable datasets are proprietary, and leading wearable foundation models trained on them are rarely open weight or include reproducible training code. To accelerate open science in wearable health, we introduce OpenMyHeartCounts (OpenMHC), the largest and most comprehensive wearable health dataset broadly accessible to qualified researchers, alongside open-source implementations and weights for recent wearable foundation models. Derived from over a decade of data collected through the My Heart Counts study app, OpenMHC comprises over 67M participant-hours of wearable recordings across up to 19 sensor channels (e.g., step count, heart rate, sleep, workouts) and up to 170 health, lifestyle, mood, and behavioral variables from 11,894 participants. Furthermore, we establish a benchmark for standardized evaluation across three tasks: (1) health and behavioral outcome prediction, (2) multivariate data imputation, and (3) time-series forecasting. Benchmarking classical methods alongside recent wearable and time-series foundation models, we find that while foundation models perform strongly, carefully tuned classical baselines like XGBoost remain highly competitive. By uniting broad-access wearable data with open code and model weights, OpenMHC provides a framework needed to drive reproducible advances in wearable health AI.

## The Economics of AI Foundation Models: Openness, Competition, and Governance

DevFeed: [The Economics of AI Foundation Models: Openness, Competition, and Governance](<https://devfeed.tech/articles/the-economics-of-ai-foundation-models-openness-competition-and-governance-139749.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2510.15200>)

Author: Fasheng Xu, Xiaoyu Wang, Wei Chen, Karen Xie

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>)

Topics: [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-foundation-models](<https://devfeed.tech/tags/ai-foundation-models.md>), [ai-policy](<https://devfeed.tech/tags/ai-policy.md>), [competition](<https://devfeed.tech/tags/competition.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [econ-th](<https://devfeed.tech/tags/econ-th.md>), [economics](<https://devfeed.tech/tags/economics.md>), [foundation-models](<https://devfeed.tech/tags/foundation-models.md>), [governance](<https://devfeed.tech/tags/governance.md>), [pricing](<https://devfeed.tech/tags/pricing.md>)

### AI overview

A two-period model finds that greater model openness can lower adaptation costs while making rivals easier to use, and usage builds model-specific experience that favors continued use of the incumbent. As learning strengthens, the incumbent's optimal openness first falls and then rises: it restricts openness to retain the deployer, then may lower its fee and rely on accumulated experience to keep the deployer despite greater openness. A full-openness mandate can disrupt that low-fee strategy, potentially reducing first-period adaptation enough to outweigh easier later entry; portability of experience across models can remove this openness trap. Integration removes the incumbent's fee but commits deployment to its model, benefiting consumers and welfare when learning is strong and potentially hurting them when it is weak. Usage subsidies can help, but may prompt larger fee increases and reduced openness, lowering adaptation and welfare even before subsidy costs.

### Source excerpt

arXiv:2510.15200v2 Announce Type: replace-cross Abstract: Foundation model developers choose how open to make their models. Greater openness lowers downstream adaptation costs but can also make competing models easier to use. In a two-period model, an incumbent chooses its license fee and openness, a deployer adapts the model for users, and an entrant offers a competing model. Usage builds model-specific experience that lowers the deployer's future adaptation cost, creating a data flywheel that favors continued use of the incumbent. We identify conditions under which optimal openness falls and then rises as learning strengthens. The incumbent moves from full openness and a high fee to restricting openness for retention, then to a low fee that builds enough usage and experience to retain the deployer despite greater openness. A full-openness mandate can induce the incumbent to abandon this low-fee strategy. The higher fee can reduce first-period adaptation despite greater openness. An openness trap arises when this welfare loss exceeds the second-period gain from easier adaptation of the entrant's model. The trap can disappear when experience transfers across models. Full integration eliminates the fee between the incumbent and deployer but commits deployment to the incumbent, potentially forgoing a rival that is cheaper to adapt. Integration raises their joint profit, consumer surplus, and welfare when learning is sufficiently strong, but can lower them when learning is weak. A small usage subsidy can benefit both developers, the deployer, and consumers. It can instead induce a fee increase larger than the subsidy and reduced openness, raising the deployer's net payment per use and adaptation costs. The resulting reduction in adaptation lowers welfare even before subsidy expenditure is deducted. These findings connect AI policy to developers' pricing and openness responses and the portability of deployment experience.

## NVIDIA Kumo Tabular: Pushing the Accuracy$\\unicode{8211}$Efficiency Frontier for Tabular Prediction

DevFeed: [NVIDIA Kumo Tabular: Pushing the Accuracy$\\unicode{8211}$Efficiency Frontier for Tabular Prediction](<https://devfeed.tech/articles/nvidia-kumo-tabular-pushing-the-accuracy-unicode-8211-efficiency-frontier-for-tabular-prediction-141830.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.23001>)

Author: Jingang Qu, Valter Hudovernik, Martin Jurkovi\\v{c}, Cedric Lorenz, David Holzm\\"uller, Federico Lopez, Dmitry Gordeev, Akihiro Nitta, Ramona Bendias, Ardrian Wong, Vedaant Jain, Vignesh Kothapalli, Gi

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Inference](<https://devfeed.tech/topics/inference.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [apache](<https://devfeed.tech/tags/apache.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [family](<https://devfeed.tech/tags/family.md>), [faster](<https://devfeed.tech/tags/faster.md>), [images](<https://devfeed.tech/tags/images.md>), [in-context-learning](<https://devfeed.tech/tags/in-context-learning.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [probabilistic](<https://devfeed.tech/tags/probabilistic.md>), [regression](<https://devfeed.tech/tags/regression.md>), [speed](<https://devfeed.tech/tags/speed.md>), [strings](<https://devfeed.tech/tags/strings.md>), [text](<https://devfeed.tech/tags/text.md>), [training-stability](<https://devfeed.tech/tags/training-stability.md>), [transformer](<https://devfeed.tech/tags/transformer.md>)

### AI overview

Kumo Tabular is a family of classification and regression models that predicts labels for new tables without task-specific training. It combines a row encoder with an in-context transformer and is pretrained on artificial tables from the Kumo prior. Its three sizes range from 27M to 214M parameters; accuracy improved with size in all tested comparisons. The abstract reports that Kumo Tabular-L ranked first on eight benchmarks, spanning probabilistic regression, very large or wide tables, and tables with strings, text, or images. All three sizes were on the accuracy-speed frontiers of TabArena and BeyondArena; the large model was reported 36 times faster than TabArena runner-up LimiX-2 and 1.7 times faster than BeyondArena runner-up TabPFN-3.5. The supplied text is an abstract, so it does not provide full methods or benchmark details. Weights are released under OpenMDW-1.1 and pretraining code under Apache 2.0.

### Source excerpt

arXiv:2611.23001v1 Announce Type: new Abstract: We introduce Kumo Tabular, a family of open tabular foundation models for classification and regression, in three sizes: Kumo Tabular-S (27M parameters), Kumo Tabular-M (61M), and Kumo Tabular-L (214M). Pretrained only on artificial tables from the Kumo prior, they can make predictions on new tables without any training. In our approach, a table encoder compresses each row into an embedding, and an in-context learning transformer predicts the unlabeled rows from the labeled ones. Kumo Tabular introduces architectural innovations that improve both training stability and predictive performance, enabling the model to scale up to 150M cells on modern GPUs. Kumo Tabular-L ranks first on eight benchmarks: TabArena, TALENT, BeyondArena, ScoringBench, STRABLE, MulTaBench, RamanBench, and TabBench-Bio. These cover probabilistic regression, tables with over 1M rows or 100K columns, and tables with strings, text, or images. Accuracy improves with model size in all tested comparisons. All three sizes lie on the accuracy-speed frontiers of TabArena and BeyondArena. Kumo Tabular-L predicts faster than the runners-up, LimiX-2 on TabArena (36 times) and TabPFN-3.5 on BeyondArena (1.7 times). Weights and pretraining code are released under OpenMDW-1.1 and Apache 2.0, respectively.

## 180B Parameters, 32GB of RAM: Pushing MoE Inference to the Limit with Transformers

DevFeed: [180B Parameters, 32GB of RAM: Pushing MoE Inference to the Limit with Transformers](<https://devfeed.tech/articles/180b-parameters-32gb-of-ram-pushing-moe-inference-to-the-limit-with-transformers-143010.md>)

Original publisher: [Read original article](<https://huggingface.co/blog/moe-expert-streaming>)

Author: Tarek Ziadé

Published: 2026-11-18T00:00:00Z

Content type: article

Language: en

Sources: [Hugging Face](<https://devfeed.tech/sources/hugging-face-blog.md>)

Topics: [Local AI](<https://devfeed.tech/topics/local-ai.md>), [Transformers](<https://devfeed.tech/topics/transformers.md>), [On-device AI](<https://devfeed.tech/topics/on-device-ai.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [quantization](<https://devfeed.tech/topics/quantization.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [llama.cpp](<https://devfeed.tech/topics/llama-cpp.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [apple](<https://devfeed.tech/tags/apple.md>), [apple-silicon](<https://devfeed.tech/tags/apple-silicon.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [bounded](<https://devfeed.tech/tags/bounded.md>), [cache](<https://devfeed.tech/tags/cache.md>), [inference](<https://devfeed.tech/tags/inference.md>), [moe](<https://devfeed.tech/tags/moe.md>), [ram](<https://devfeed.tech/tags/ram.md>), [transformers](<https://devfeed.tech/tags/transformers.md>)

### AI overview

The excerpt reports a Transformers prototype that streams only router-selected MoE experts from SSD through a bounded device cache, remaps expert IDs to shared cache slots, and reuses `grouped_mm` for expert computation. On a 32 GB M5, after quantizing experts to 3 bits, compressing dense layers, and adding a separate disk path for Qwen3.8-Flash-Next's n-gram table, the system reached about 14 tokens/s--roughly 6% behind Strata on short decode, with slightly faster prefill. It uses about 18 GiB after loading; a 6K prompt peaks at 25.4 GiB, and the author reports needing to close the browser to avoid swapping. The remaining gap is attributed mainly to host round trips and GPU idle time between layers; closing it would require deeper MPS synchronization control unavailable in PyTorch. The API is still under discussion, and the approach currently supports models with the standard experts layout.

### Source excerpt

We're on a journey to advance and democratize artificial intelligence through open source and open science.

## Nvidia says its open-source AI strategy remains intact after Jensen Huang's exit

DevFeed: [Nvidia says its open-source AI strategy remains intact after Jensen Huang's exit](<https://devfeed.tech/articles/nvidia-says-its-open-source-ai-strategy-remains-intact-after-jensen-huang-s-exit-138647.md>)

Original publisher: [Read original article](<https://techcrunch.com/2026/11/17/nvidia-says-its-open-source-ai-strategy-remains-intact-after-jensen-huangs-exit/>)

Author: Sarah Perez

Published: 2026-11-17T23:04:35Z

Content type: news

Language: en

Sources: [TechCrunch | AI News & Artificial Intelligence](<https://devfeed.tech/sources/techcrunch-ai-news-artificial-intelligence.md>)

Topics: [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Nemotron](<https://devfeed.tech/topics/nemotron.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>)

Tags: [acquisition](<https://devfeed.tech/tags/acquisition.md>), [adoption](<https://devfeed.tech/tags/adoption.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agentic-ai-systems](<https://devfeed.tech/tags/agentic-ai-systems.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-adoption](<https://devfeed.tech/tags/ai-adoption.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [ai-systems](<https://devfeed.tech/tags/ai-systems.md>), [ai-technology](<https://devfeed.tech/tags/ai-technology.md>), [alice-huang](<https://devfeed.tech/tags/alice-huang.md>), [analysts](<https://devfeed.tech/tags/analysts.md>), [ceo](<https://devfeed.tech/tags/ceo.md>), [chinese-spy](<https://devfeed.tech/tags/chinese-spy.md>), [hugging-face](<https://devfeed.tech/tags/hugging-face.md>), [interim-ceo](<https://devfeed.tech/tags/interim-ceo.md>), [jensen-huang](<https://devfeed.tech/tags/jensen-huang.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [open-source-ai](<https://devfeed.tech/tags/open-source-ai.md>), [tc](<https://devfeed.tech/tags/tc.md>)

### AI overview

Nvidia interim CEO Colette Kress said the company's open-model strategy and planned $12.9 billion acquisition of Hugging Face remain on track after Jensen Huang's departure. She cited a 35-fold year-over-year rise in weekly token volume on OpenRouter, including 3.5-fold growth over three months, as evidence of accelerating adoption. Kress also described Nemotron as OpenRouter's most widely used U.S. open model. The acquisition was not addressed directly in the earnings call's Q&A, so the excerpt supports Nvidia's stated plans but provides no further transaction detail.

### Source excerpt

Nvidia's interim CEO reaffirmed its open-source AI strategy and plans to acquire Hugging Face, despite Jensen Huang's sudden departure.

## Read the EmbeddingGemma 2 technical report.

DevFeed: [Read the EmbeddingGemma 2 technical report.](<https://devfeed.tech/articles/read-the-embeddinggemma-2-technical-report-138184.md>)

Original publisher: [Read original article](<https://blog.google/innovation-and-ai/technology/developers-tools/embeddinggemma-2-technical-report/>)

Author: The EmbeddingGemma 2 team

Published: 2026-11-17T17:00:00Z

Content type: release

Language: en

Sources: [Developer tools](<https://devfeed.tech/sources/developer-tools.md>)

Topics: [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [model architecture](<https://devfeed.tech/topics/model-architecture.md>), [LLM Techniques](<https://devfeed.tech/topics/llm-techniques.md>)

Tags: [audio](<https://devfeed.tech/tags/audio.md>), [developer-tools](<https://devfeed.tech/tags/developer-tools.md>), [embeddinggemma-2](<https://devfeed.tech/tags/embeddinggemma-2.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [experimental](<https://devfeed.tech/tags/experimental.md>), [explainer](<https://devfeed.tech/tags/explainer.md>), [gemini](<https://devfeed.tech/tags/gemini.md>), [gemma-4](<https://devfeed.tech/tags/gemma-4.md>), [images](<https://devfeed.tech/tags/images.md>), [model](<https://devfeed.tech/tags/model.md>), [multimodal](<https://devfeed.tech/tags/multimodal.md>), [multimodal-model](<https://devfeed.tech/tags/multimodal-model.md>), [none](<https://devfeed.tech/tags/none.md>), [report](<https://devfeed.tech/tags/report.md>), [space](<https://devfeed.tech/tags/space.md>), [technical](<https://devfeed.tech/tags/technical.md>), [training](<https://devfeed.tech/tags/training.md>), [vector](<https://devfeed.tech/tags/vector.md>), [video](<https://devfeed.tech/tags/video.md>)

### AI overview

EmbeddingGemma 2 extends Google's open text-only embedding model to map text, images, audio, video, and combinations of these into a unified embedding space, with a more compact model backbone. The report covers its training strategy, evaluation methodology, and experimental results. The excerpt gives no specifics about those methods or results.

### Source excerpt

The Google EmbeddingGemma 2 technical report details our new multimodal model, covering its unified vector space and compact design. Read the report.

## The Complete Guide to HIPAA/GDPR-Compliant AI Inference: Self-Hosting Open-Source Models

DevFeed: [The Complete Guide to HIPAA/GDPR-Compliant AI Inference: Self-Hosting Open-Source Models](<https://devfeed.tech/articles/the-complete-guide-to-hipaa-gdpr-compliant-ai-inference-self-hosting-open-source-models-138036.md>)

Original publisher: [Read original article](<https://hackernoon.com/the-complete-guide-to-hipaagdpr-compliant-ai-inference-self-hosting-open-source-models?source=rss>)

Author: SIEmon

Published: 2026-11-17T14:28:59Z

Content type: article

Language: en

Sources: [HackerNoon](<https://devfeed.tech/sources/hackernoon.md>)

Topics: [Self Hosted AI Tools](<https://devfeed.tech/topics/self-hosted-ai-tools.md>), [LLM security](<https://devfeed.tech/topics/llm-security.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Self-hosted](<https://devfeed.tech/topics/self-hosted.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-inference](<https://devfeed.tech/tags/ai-inference.md>), [ai-security](<https://devfeed.tech/tags/ai-security.md>), [air-gapped](<https://devfeed.tech/tags/air-gapped.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [apache-2-0](<https://devfeed.tech/tags/apache-2-0.md>), [api](<https://devfeed.tech/tags/api.md>), [apis](<https://devfeed.tech/tags/apis.md>), [audit-trail](<https://devfeed.tech/tags/audit-trail.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [gdpr](<https://devfeed.tech/tags/gdpr.md>), [gdpr-compliance](<https://devfeed.tech/tags/gdpr-compliance.md>), [good-company](<https://devfeed.tech/tags/good-company.md>), [hipaa](<https://devfeed.tech/tags/hipaa.md>), [hipaa-compliance](<https://devfeed.tech/tags/hipaa-compliance.md>), [hipaa-compliant-ai-inference](<https://devfeed.tech/tags/hipaa-compliant-ai-inference.md>), [hosted](<https://devfeed.tech/tags/hosted.md>), [hosting](<https://devfeed.tech/tags/hosting.md>), [inference](<https://devfeed.tech/tags/inference.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [open-source-llms](<https://devfeed.tech/tags/open-source-llms.md>), [privacy](<https://devfeed.tech/tags/privacy.md>), [self-hosted-inference-servers](<https://devfeed.tech/tags/self-hosted-inference-servers.md>), [self-hosted-llm](<https://devfeed.tech/tags/self-hosted-llm.md>), [self-hosting](<https://devfeed.tech/tags/self-hosting.md>), [servers](<https://devfeed.tech/tags/servers.md>)

### AI overview

HIPAA and GDPR do not require self-hosting: the excerpt says OpenAI and Anthropic offer HIPAA-related contracts for eligible features. Hosted APIs may suit low or uneven usage or a need for a frontier model; self-hosting may fit when contract scope, retention, data location, version pinning, vendor count, or steady high volume makes hosted service impractical. Agents increase exposure because one query can trigger many model calls. The checklist recommends eligible cloud services with a BAA or DPA, risk analysis and (for GDPR) a DPIA, private authenticated endpoints, MFA, metadata-only logs, telemetry review, redaction, retention and deletion rules, pinned model versions, and an inventory of tools and MCP servers. SIE is described as an Apache 2.0 server deployable in the customer's cloud, including air-gapped operation; however, Superlinked does not currently sign BAAs, so the excerpt recommends keeping vendor support away from PHI.

### Source excerpt

HIPAA and GDPR compliance for AI inference: compare hosted APIs and self-hosting, explore privacy risks and secure open-source inference servers.

[Next page](<https://devfeed.tech/topics/open-source-models-datasets.md?cursor=WyIyMDI2LTExLTE3VDE0OjI4OjU5KzAwOjAwIiwgIjAwYjA5MzI0LWY4YjMtNGU2Ni1hNGFlLWFiYmIyYzE0ZjIwMiJd>)