# Gpt 3 Published articles for Gpt 3. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Evaluating gpt-4o-mini vs. gpt-3.5-turbo for RAG applications DevFeed: [Evaluating gpt-4o-mini vs. gpt-3.5-turbo for RAG applications]() Original publisher: [Read original article]() Author: Pamela Fox (noreply@blogger.com) Published: 2025-03-06T08:22:00Z Content type: comparison Language: en Sources: [Pamela Fox]() Topics: [Retrieval Augmented Generation (RAG)](), [Artificial Intelligence](), [Azure]() Tags: [ai](), [azure](), [cost](), [evaluation](), [gpt](), [gpt-3](), [latency](), [openai](), [rag](), [vs]() ### AI overview The article evaluates gpt-4o-mini against gpt-35-turbo as the default model for the Azure RAG sample application. Across evaluations using HR documents and the author's blog, gpt-4o-mini had comparable groundedness and relevance, produced longer answers, took more generation time, and had lower overall cost because of its lower per-token pricing. ### Source excerpt The azure-search-openai-demo repository was first created in March 2023 and is now the most popular RAG sample solution for Azure. Since the world of generative AI changes so rapidly, we've made many upgrades to its underlying packages and technologies over the past two years. But we've never changed the default GPT model used for the RAG flow: gpt-35-turbo. Why, when there are new models that are cheaper and reportedly better, such as gpt-4o-mini? Well, changing the model is one of the most significant changes you can make to impact RAG answer quality, and I did not want to make the change without thorough evaluation. Good news! I have now run several bulk evaluations on different RAG knowledge bases, and I feel fairly confident that a switch to gpt-4o-mini is a positive overall change, with some caveats. In my evaluations, gpt-4o-mini generates answers with comparable groundedness and relevance. The time-per-token is slightly less, but the answers are 50% longer on average, thus they take 45% more time for generation. The additional answer length often provides additional details based off the context, especially for questions where the answer is a list or a sequential process. The gpt-4o-mini per-token pricing is about 1/3 of gpt-35-turbo pricing, which works out to a lower overall cost. Let's dig into the results more in this post. Evaluation results I ran bulk evaluations on two knowledge bases, starting with the sample data that we include in the repository, a bunch of invented HR documents for a fictitious company. Then, since I always like to evaluate knowledge that I know deeply, I also ran evaluations on a search index composed entirely of my own blog posts from this very blog. Here are the results for the HR documents, for 50 Q/A pairs: metric stat gpt-35-turbo gpt-4o-mini gpt_groundedness pass_rate 0.98 0.98 mean_rating 4.94 4.9 gpt_relevance pass_rate 0.98 0.96 mean_rating 4.42 4.54 answer_length mean 667.7 934.36 latency mean 2.96 3.8 citations_matched ## SudoLang: A Pseudocode Language for Prompting LLMs DevFeed: [SudoLang: A Pseudocode Language for Prompting LLMs]() Original publisher: [Read original article]() Author: Eric Elliott Published: 2023-04-01T02:15:41Z Content type: article Language: en Sources: [Eric Elliot]() Topics: [Large Language Model](), [Programming](), [GitHub Copilot](), [GitHub]() Tags: [ai](), [chatgpt](), [coding](), [context](), [copilot](), [developer](), [gpt-3](), [hallucinations](), [language](), [language-models](), [llms](), [nlp](), [programming](), [programming-language](), [software-development](), [sudolang](), [technology]() ### AI overview The article introduces SudoLang, a formalized pseudocode language intended to help prompt large language models. It discusses how GPT-4 improved the usefulness of detailed pseudocode instructions while noting that LLM outputs can still contain errors and hallucinations. ### Source excerpt Pseudocode is a fantastic way to sketch programs using informal, natural language, without worrying about specific syntax. It's like sketching your thoughts before diving into the nitty-gritty of coding. It's useful for brainstorming and communicating ideas with others. I have been using pseudocode to express ideas to Large Language Models (LLMs) since GPT-3 was announced in mid 2020. However, up until GPT-4 was released, it didn't work extremely well. It sometimes worked, but I threw away most completions on GPT-3 and GPT-3.5. Even so, LLMs were already extremely useful in that state. In a 2022 study, GitHub found that Copilot shaved 55% of the time off a project task assigned to 95 people: 45 using Copilot, and the rest without. In other words, even before LLM pseudocode languages like SudoLang, LLMs were already making a huge impact on developer productivity. But GPT-4 raises the bar. Not a little bit, but a lot. First, let's address the obvious. GPT-4 is: 40% less likely to deliver an incorrect response. (But still prone to errors and hallucinations. Be sure to check its output!) Multi-modal and can interact with both text and images. (I haven't had the chance to test this, yet.) Sweeping intelligence and training improvements. (This may be an understatement. The improvement is astonishing.) Much longer context available. (It can help you build much larger programs!) GPT-4 scores significantly better on a broad range of human level knowledge and intelligence assessment tests, including the Bar, LSAT, SAT, GRE, and AP placement tests. (Hot tip: GPT-4 is still not a lawyer even if it can pass the Bar. It will still sometimes hallucinate facts that aren't true. Please don't let it get you into legal trouble!) Source: GPT-4 Technical Report The really interesting upshot of all this is that GPT-4 now has the ability to follow detailed, step by step instructions, even through complex abstractions and long chains of function compositions. I've been playing with a lot o ## Supabase Clippy: ChatGPT for Supabase Docs DevFeed: [Supabase Clippy: ChatGPT for Supabase Docs]() Original publisher: [Read original article]() Author: Paul Copplestone Published: 2023-02-07T07:00:00Z Content type: tutorial Language: en Sources: [Supabase Blog]() Topics: [AI Chat]() Tags: [chatgpt](), [developer-experience](), [docs](), [embeddings](), [gpt-3](), [knowledge-base](), [openai](), [postgres](), [search](), [streaming]() ### AI overview Supabase introduces an MVP ChatGPT-style interface for its documentation. It retrieves relevant documentation sections using OpenAI embeddings stored in Postgres with pgvector, then supplies them as context for GPT-3 responses. ### Source excerpt Creating a ChatGPT interface for the Supabase documentation. ## GPT-3's improved capabilities and implications for application security DevFeed: [GPT-3's improved capabilities and implications for application security]() Original publisher: [Read original article]() Author: Adam Published: 2022-12-09T00:00:00Z Content type: opinion Language: en Sources: [Shostack & Friends Blog]() Topics: [OpenAI](), [Artificial Intelligence](), [Language models](), [Application Security](), [prompt]() Tags: [ai](), [application-security](), [chatgpt](), [gpt](), [gpt-3](), [language-models](), [openai](), [prompt]() ### AI overview The author reflects on GPT-3's marked improvement, its ability to produce reasonably convincing text, and the implications for application security and public information. ### Source excerpt The OpenAI chatbot is shockingly improved -- its capabilities deserve attention. ## GPT-3 DevFeed: [GPT-3]() Original publisher: [Read original article]() Author: Adam Published: 2022-12-09T00:00:00Z Content type: article Language: en Sources: [Shostack & Friends Blog]() Topics: [ChatGPT](), [Library](), [spoofing](), [Back end](), [Kubernetes]() Tags: [backend](), [chatgpt](), [client](), [gpt](), [gpt-3](), [kubernetes](), [libraries](), [library](), [service](), [spoofing]() ### AI overview The article reports that GPT-3 was used to generate sample STRIDE threat libraries for backend-to-backend services in Kubernetes and client-side threats. It notes that the output may be imperfect and requires human judgment. ### Source excerpt Text captured from GPT-3 ## AI and poetry meet Trino DevFeed: [AI and poetry meet Trino]() Original publisher: [Read original article]() Author: Martin Traverso, Marcos Traverso Published: 2022-12-06T00:00:00Z Content type: opinion Language: en Sources: [Trino Blog]() Topics: [Artificial Intelligence](), [Poetry](), [OpenAI](), [data](), [Code](), [real-time](), [systems]() Tags: [ai](), [blog](), [blog-post](), [code](), [gpt](), [gpt-3](), [openai](), [poetry](), [prompt]() ### AI overview The article describes an experiment using GPT-3 to generate poetry about Trino. It presents two generated poems that highlight Trino's data integration, SQL, real-time analytics, distributed-system, scalability, and data-source capabilities, then invites readers to try the experiment themselves. ### Source excerpt Sometimes when working in the depth of the query engine core, the planner, the optimizer, or some other tricky problem or code, you run out of steam and need a distraction. Your brain will quietly work on the problem in the background while you get to read or do something interesting. During one break we found out more about GPT-3 and the recent improvements on the OpenAI project. So we had to try. ## Our product got free advertising from AI DevFeed: [Our product got free advertising from AI]() Original publisher: [Read original article]() Author: Glauber Costa Published: 2022-10-18T00:00:00Z Content type: article Language: en Sources: [Turso Blog]() Topics: [Turso](), [SQLite](), [libSQL]() Tags: [chiselstrike](), [gpt-3](), [libsql](), [product](), [serendipity](), [sqlite](), [technology](), [turso]() ### AI overview The article discusses Turso's product receiving free advertising from AI. The supplied text notes that the post references an older version of Turso, describes Turso as a ground-up rewrite of SQLite, and explains Turso Cloud's database access, synchronization, offline support, and file-based architecture. ### Source excerpt An entrepreneur's unlikely journey from environmental concerns through artificial intelligence leads him to the ChiselStrike BaaS. ## Transformers for software engineers DevFeed: [Transformers for software engineers]() Original publisher: [Read original article]() Author: Nelson Elhage Published: 2022-04-01T20:00:00Z Content type: tutorial Language: en Sources: [Nelson Elhage]() Topics: [Transformer architecture](), [Programming](), [Software Engineering](), [Reverse Engineering](), [Transformers]() Tags: [anthropic](), [architecture](), [codex](), [deep-learning](), [gpt-3](), [model-architecture](), [programming](), [reverse-engineering](), [software-engineering](), [storm](), [transformer-architecture](), [transformers]() ### AI overview This tutorial explains Transformer architecture for software engineers, using software engineering and programming perspectives to discuss how GPT-style Transformer models work. It also connects the architecture to interpretability and reverse-engineering efforts. ### Source excerpt Ever since its introduction in the 2017 paper, Attention is All You Need, the Transformer model architecture has taken the deep-learning world by storm. Initially introduced for machine translation, it has become the tool of choice for a wide range of domains, including text, audio, video, and others. Transformers have also driven most of the massive increases in model scale and capability in the last few years. OpenAI's GPT-3 and Codex models are Transformers, as are DeepMind's Gopher models and many others. ## Автоматическая многоязычная модерация сообщений для миллионов пользователей DevFeed: [Автоматическая многоязычная модерация сообщений для миллионов пользователей]() Original publisher: [Read original article]() Author: punkerpunker (Badoo) Published: 2022-02-01T10:50:28Z Content type: article Language: ru Sources: [Badoo EN](), [Badoo RU]() Topics: [Artificial Intelligence](), [Natural language processing](), [Large Language Model]() Tags: [ai](), [badoo](), [bert](), [data-science](), [gpt](), [gpt-3](), [natural-language-processing](), [nlp](), [tag-055aee430837](), [tag-13e2af703774](), [tag-1cd610c0e518](), [tag-5cb789e9aaa3](), [tag-9d8cf70dc46c](), [tag-e69b694e3ce2]() ### AI overview The article describes Badoo and Bumble's Rude Message Detector, a machine-learning system for multilingual moderation of rude messages. It discusses the system's architecture, deployment infrastructure, implementation challenges, and results, including the use of transformer-based architectures and large language models such as BERT, GPT-2, and GPT-3. ### Source excerpt Как мы все знаем, слова обладают силой. Ими можно вдохновить людей, но можно и навредить. Мы в Badoo и Bumble стараемся оградить пользователей от неприятных ситуаций, поэтому внедрили инструмент Rude Message Detector. Это многоязычный детектор грубых высказываний, работающий на основе машинного обучения. В этой статье мы расскажем о технических особенностях нашего решения, сложностях, с которыми столкнулись в ходе реализации проекта, и его результатах. Читать далее