# HaloGuard 2.0: An Open-Weights Constitutional Output Guard for Multilingual Response Safety

DevFeed: [HaloGuard 2.0: An Open-Weights Constitutional Output Guard for Multilingual Response Safety](<https://devfeed.tech/articles/haloguard-2-0-an-open-weights-constitutional-output-guard-for-multilingual-response-safety-114068.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.02134>)

Author: Ashmiya Lenin, Navaneeth Sangameswaran, Preetham S

Published: 2026-11-03T05:00:00Z

Content type: release

Language: en

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

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

Tags: [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [concurrent](<https://devfeed.tech/tags/concurrent.md>), [costs](<https://devfeed.tech/tags/costs.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-cr](<https://devfeed.tech/tags/cs-cr.md>), [false-positives](<https://devfeed.tech/tags/false-positives.md>), [input](<https://devfeed.tech/tags/input.md>), [jailbreak](<https://devfeed.tech/tags/jailbreak.md>), [languages](<https://devfeed.tech/tags/languages.md>), [open-weights](<https://devfeed.tech/tags/open-weights.md>), [operational](<https://devfeed.tech/tags/operational.md>), [red-teaming](<https://devfeed.tech/tags/red-teaming.md>), [responses](<https://devfeed.tech/tags/responses.md>), [safety](<https://devfeed.tech/tags/safety.md>), [stop](<https://devfeed.tech/tags/stop.md>), [stream](<https://devfeed.tech/tags/stream.md>), [token](<https://devfeed.tech/tags/token.md>), [training](<https://devfeed.tech/tags/training.md>)

## AI overview

HaloGuard 2.0 is an open-weight streaming output guard that scores each response token in one forward pass, allowing a deployment to stop generation when harmful content is confirmed. Its first training setup learned to judge prompts instead of responses, correctly handling only 13.9% of matched safe/unsafe pairs; a corrective phase on 106,043 counterfactual responses addressed this, and the released model averages the base and corrective models. On six English benchmarks, the 0.8B model reaches 83.6 macro F1, up from 72.8, with 38% fewer false positives; it is reported to match leading full-response guards with 5-10 times fewer parameters and score 2.6 points above Qwen3Guard-Stream-4B. The multilingual 0.8B variant scores 83.1 F1 across 17 languages, 5.0 points above the strongest baseline. At 64 concurrent streams on one H100, reported cost is 0.37 ms per token per stream. The excerpt also says both guards undergo weekly red-teaming.

## Source excerpt

arXiv:2611.02134v1 Announce Type: cross Abstract: An input guard decides whether a request reaches the model, but never sees what the model says. Harm that emerges during generation, because a jailbreak passed the input layer or a benign request drifted into operational detail, is visible only on the output side and only as tokens arrive. We present HaloGuard 2.0, an open-weight streaming output guard that assigns a safe/unsafe probability to every response token in a single forward pass, so a deployment can stop the stream once harm is confirmed. The guard reuses the constitution of HaloGuard 1.0, our input guard, to organise a 478,532-response training corpus, and supervises each token from the point where the response becomes harmful. We find that a guard trained on conventionally shaped data learns a shortcut specific to streaming: it judges the prompt instead of the response, and handles only 13.9% of matched safe/unsafe response pairs correctly. We correct this with a second phase on 106,043 counterfactual responses built around the base guard's own failure modes, and release the equal-weight average of the base and corrective models. On six English response-safety benchmarks, HaloGuard 2.0-0.8B reaches a macro F1 of 83.6 (83.3 on held-out halves of the decision-rule search), up from 72.8, with 38% fewer false positives. It matches the strongest full-response guards with 5-10x fewer parameters and is 2.6 points above the strongest streaming guard, Qwen3Guard-Stream-4B. A multilingual variant, HaloGuard 2.0-Poly-0.8B, trained on PolyGuardMix, reaches 83.1 F1 on the response labels of PolyGuardPrompts across 17 languages, 5.0 points above the strongest baseline. On one H100 the guard costs 0.37 ms per token per stream when moderating 64 concurrent responses. Both guards are hardened continuously through the weekly red-teaming protocol of HaloGuard 1.0, extended here to the output layer.