Check out our latest SAIL blog post on HalluWorld! This benchmark puts LLMs in fully specified worlds to automatically measure not just whether models hallucinate, but why:
32227.8K
Stanford AI Lab✓
@StanfordAILab
The Stanford Artificial Intelligence Laboratory (SAIL), a leading #AI lab since 1963. ⛵️🤖 Emmy-winning video: https://t.co/lV9smZTC1m
337 Following277K Followers3.6K Posts
- Stanford AI LabRT @bingao_chen: Wrist cameras give robots close-up views, but only where the hand happens to be. What if the policy could decide where to look? Introducing FoveaVLA: a VLA that replaces the wrist camera with learned foveation actions. FoveaVLA predicts where it will need to look next, then crops that region from the full-resolution external image. The full view gives context; the foveal crop gives close-up detail. With one external camera and no wrist camera, FoveaVLA reaches 70.8% success on 48 real-world trials, vs. 52.1% for π0.5 with a wrist camera. • No additional demonstrations • No camera calibration • No wrist camera Paper, code, and models are all open 🧵0290639
- Stanford AI LabRT @AriKagan_: quite the statement from Anthropic's head of security
Jason D. Clinton 🔸 @JasonDClinton@hinapansari yes. it was always going to be this way. we tried to warn the world.09510578 
Posts @hinapansari@hinapansari yes. it was always going to be this way. we tried to warn the world.7041081.3K1.3M- Stanford AI LabRT @mjnblack: Things are moving so fast that everyone seems to have completely missed how insane it is that AI companies have been allowed to unleash agentic AI products with no regulatory guidelines.Madison Mills @MadisonMills22SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios. The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios. Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said. Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior. Read my latest for Axios here:02K0631
SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios. The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios. Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said. Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior. Read my latest for Axios here:1.4K6.9K19.1K9.1M- Stanford AI LabRT @SISLaboratory: Now online! Recording of Marc Schlichting’s PhD defense (Nov 15, 2025) on generative models for safety in autonomous systems, from failure discovery in a 1200-dim F-16 system to an LLM decision-support tool for aviation emergencies tested in real flights.030513
- Stanford AI LabRT @drmapavone: Do Vision-Language-Action models mean what they say? Embodied chain-of-thought offers a promising way to leverage additional inference-time computation for robot policies, with empirical evidence that it can improve downstream task performance. However, in practice the extent to which this chain-of-thought actually influences a robot policy’s chosen actions is not fully understood yet. In our new paper on #embodied #reasoning, we formally define the notion of faithfulness i.e., whether the chain-of-thought truthfully reflects the policy's decision for #Physical #AI and argue that the faithfulness of robot policies should play an explicit role in algorithm design. We train a critic of embodied faithfulness, identifying the signatures of inconsistent and incoherent reasoning, and use it as a reward signal during reinforcement learning (RL) post-training of a robot policy. We call our critic Pinocchio 🤥: a model that learns to tell when a robot policy is 'lying' about the justifications for a decision. Pinocchio scores two complementary failure modes: whether the reasoning is grounded in what the robot actually sees, and whether it is coherent step-to-step through the chain-of-thought. We find that post-training for faithfulness yields a +18% improvement in reasoning consistency over standard baselines, and critically, makes the policy 1.6× more responsive to rare, out-of-distribution scenarios that confuse SoTA models. These findings suggest that faithful embodied reasoning is not merely a post-hoc interpretability property, but an important ingredient for building more robust and trustworthy autonomous systems. 📄 Paper: 🌐 Project page: 💻 Code: Coming soon! Great work led by Matthew Foutter, in collaboration with Matteo Cercola Lena Wild Yunshan Wang Michelle Li and @DanieleGammelli @StanfordEng @StanfordAILab @StanfordHAI070500
- Stanford AI LabRT @drmapavone: Ever wondered what it’s like to learn robot autonomy @Stanford? Now you can. Today, we’re opening up the robot autonomy teaching materials that my collaborators and I have been developing over the past several years—for anyone, anywhere, completely free. At the heart of this effort is Principles of Robot Autonomy (PoRA), a textbook inspired by the Stanford robot autonomy course series I’ve been teaching since 2017. Robot autonomy has evolved from a research discipline into a technology deployed in the real world. Building capable autonomous systems requires more than understanding individual algorithms: it requires understanding how perception, state estimation, planning, control, and learning work together in a complete autonomy stack. That’s what we set out to teach with PoRA: bringing classical robotics and modern AI together in one coherent framework, grounded in both theory and practice. Coming soon in print from Cambridge University Press, Principles of Robot Autonomy is also freely available online and comes with a comprehensive collection of open educational resources: 📖 Textbook, — Foundations of robot autonomy, from classical methods to modern AI-based approaches. 🎥 Video lectures, Lectures from my Stanford courses on robot autonomy and on optimal and learning-based control. 💻 Interactive exercises, — Hands-on Python implementations and Jupyter notebooks. 📝 Teaching materials, — Lecture slides and resources for instructors. 🌐 Explore all PoRA resources, — Including Ask the Book, an AI study assistant integrated throughout the textbook to answer questions, explain derivations, and connect ideas across chapters. 📣 Read the Stanford announcement: Our goal is simple: to make high-quality education in robot autonomy accessible to students, researchers, engineers, educators, and anyone curious about how autonomous robotic systems work. I believe that sharing knowledge is fundamental to the mission of a university. Autonomous robotics—and Physical AI more broadly—are among the defining technological frontiers of our time. We want to help more people understand these technologies, build with them, and advance the field. A huge thank you to my co-authors, @DanieleGammelli, Joseph Lorenzetti, @katielulula, and @GioeleZardini, and to everyone who helped bring this project to life. Explore the resources, use them in your own learning or teaching, and please share them with anyone who might find them useful! #Robotics #PhysicalAI #ArtificialIntelligence #AutonomousSystems #Education @Stanford @StanfordEng @StanfordAILab @StanfordHAI01970457
- Stanford AI LabRT @Anikait_Singh_: LLMs deployed in specialized domains, from enterprise codebases to law firms, dig through a mountain of unstructured context! Our new paper, Librarian Agents, asks: can an agent adapt by curating its own long-context workspace, anticipating queries before they arrive? 🧵⬇️0320532
- Stanford AI LabRT @StanfordHAI: 7 years ago: Stanford HAI officially launched. The institute opened in March 2019 with a focus on advancing AI research, education, policy, and practice to improve the human condition, drawing on faculty and students from all seven Stanford schools. #HAIOverTheYears040450
- Token-Gated LLMs achieve 11x faster processing over long-context inputs by learning to skim context and only forward tokens that matter. We saw this paradigm shift before when going from dense models to MoEs, and now TokenGate enables another generation of efficient LLMs!Liana @lianapatel_Introducing ✨Token-Gated Language Models✨ for up to 11x faster prefill by learning to skim context, conditionally forwarding only tokens that matter. Akin to the shift from dense models to MoEs, the next frontier of efficient LLMs is conditional computation over a new scaling axis: context tokens🧵👇91913210.7K
Introducing ✨Token-Gated Language Models✨ for up to 11x faster prefill by learning to skim context, conditionally forwarding only tokens that matter. Akin to the shift from dense models to MoEs, the next frontier of efficient LLMs is conditional computation over a new scaling axis: context tokens🧵👇
296436044.1K- Stanford AI LabRT @james_y_zou: Why stick with one agent harness for the entire task? Surprisingly, switching harnesses mid-run (e.g. Codex → Pi) can dramatically reduce cost/tokens while improving performance! Harness cascade opens a new Pareto frontier.Yongchan Kwon @ykwon_0407What if your agent could switch harnesses mid-run? Imagine starting with Claude Code, then switching to Pi or Codex. Would it approach the task differently? Catch mistakes it previously missed? Could different harnesses complement each other? 🧵0120565
Yongchan Kwon@ykwon_0407What if your agent could switch harnesses mid-run? Imagine starting with Claude Code, then switching to Pi or Codex. Would it approach the task differently? Catch mistakes it previously missed? Could different harnesses complement each other? 🧵8165212.3K- Stanford AI LabRT @StefanoErmon: 1/4 Today, we’re releasing Mercury 3. Mercury 3 is by far our most powerful model and the most intelligent diffusion LLM ever released. It outperforms GPT-6 Luna and Haiku 5.5 on quality while running at 1000+ tok/sec.04570490
- Stanford AI LabRT @AmrollahiD: Autoformalization is the task of translating natural language into formal, machine-checkable representations. An open question: when an LLM does this, how do we know the output is faithful to the original meaning? A small step towards answering it:030552
- Stanford AI LabRT @DuranteZane: 1/ Can AI help doctors make more informed decisions based on patient activity and behavior over the course of a hospital stay? Today we introduce TRACER: a training-free multimodal harness that enables LLMs and VLMs to track patient activity using bedside video and health records. Blogpost: Paper:0380583
- Stanford AI LabRT @agarwal_suchir: 1/8 Introducing LILIE: a Language-Informed and Lossless Image Encoder. We train a language-informed encoder to supervise end-to-end visual generative models. Against a pixel-space JiT model, LILIE improves gFD on GPIC by 14 points (66.84 -> 53.01).0270503
- Stanford AI LabRT @FrancoisChauba1: this was one of the best YCPCs we've done. huge thank you to mykel and clark. real ai safety ideas. grounded in science. not science fiction.Y Combinator @ycombinatorThis week’s Paper Club is all about AI safety. Not the headlines, but the science. AI safety has been a research field for decades, with its own conferences, methods, and a track record in systems like aircraft collision avoidance, where the standard is one catastrophic failure per billion flight hours. We'll hear from @ClarkBarrett7 and Mykel Kochenderfer (@aiprof_mykel), Stanford professors and co-directors of the Stanford Center for AI Safety. Together with @FrancoisChauba1, they break down the OpenAI–Hugging Face incident, how formal verification can mathematically prove properties of code, how aircraft safety techniques are being adapted to language models, and why AI may need its own NTSB. 00:00 - Science fiction vs. science fact 04:32 - Engineering problem, or P(doom) 06:04 - What actually happened at Hugging Face 07:53 - Where "10% in 10 years" comes from 11:21 - The real risk: open weights in the wrong hands 17:25 - Clark Barrett: when AI met automated reasoning 20:35 - CSLib: proving code correct in Lean 24:32 - From English spec to verified code in one click 27:55 - Mykel Kochenderfer: how do we know a system is safe 32:11 - Adaptive stress testing, from aircraft to LLMs 36:57 - Verifying neural networks 40:30 - Runtime monitoring and BetterBench 43:53 - Panel: reacting to the past three weeks 45:55 - An NTSB for AI 52:18 - Is cybersecurity solvable in five years 55:00 - If you were CISO of Bank of America 1:03:06 - The Swiss cheese model 1:05:43 - Audience Q&A0100613
This week’s Paper Club is all about AI safety. Not the headlines, but the science. AI safety has been a research field for decades, with its own conferences, methods, and a track record in systems like aircraft collision avoidance, where the standard is one catastrophic failure per billion flight hours. We'll hear from @ClarkBarrett7 and Mykel Kochenderfer (@aiprof_mykel), Stanford professors and co-directors of the Stanford Center for AI Safety. Together with @FrancoisChauba1, they break down the OpenAI–Hugging Face incident, how formal verification can mathematically prove properties of code, how aircraft safety techniques are being adapted to language models, and why AI may need its own NTSB. 00:00 - Science fiction vs. science fact 04:32 - Engineering problem, or P(doom) 06:04 - What actually happened at Hugging Face 07:53 - Where "10% in 10 years" comes from 11:21 - The real risk: open weights in the wrong hands 17:25 - Clark Barrett: when AI met automated reasoning 20:35 - CSLib: proving code correct in Lean 24:32 - From English spec to verified code in one click 27:55 - Mykel Kochenderfer: how do we know a system is safe 32:11 - Adaptive stress testing, from aircraft to LLMs 36:57 - Verifying neural networks 40:30 - Runtime monitoring and BetterBench 43:53 - Panel: reacting to the past three weeks 45:55 - An NTSB for AI 52:18 - Is cybersecurity solvable in five years 55:00 - If you were CISO of Bank of America 1:03:06 - The Swiss cheese model 1:05:43 - Audience Q&A343425653.7K- Stanford AI LabRT @Jiaru_Zou: Your LLM agent has to think out loud, token by token, before it can search, run code, or act in the world. We humans don’t. We ponder in silence and speak when it matters. Human language was never meant to be the limit of machine thought. Introducing 🌗Hybrid Reasoning Agents - Think in latent space, act through text - Two modes, one unified multi-turn flow - No black box, fully readable latent thoughts - Any LLM can become a HybridAgent01240633
- Check out our latest SAIL blog post on TRACER! This multimodal system uses bedside video and EHRs to help clinicians understand how patient behavior changes over a hospital stay:443412.1K
- What if language models wrote the way that we think through ideas: rough sketch first with details last? New work by @KumbongHermann, @ishanskhare, and team introduces Next-Concept Modeling (NCM), a parallel generation framework achieving comparable or better quality than strong diffusion baselines while being 10-31× faster⚡️across unconditional generation, summarization & translation, and downstream reasoning tasks.Hermann Kumbong @KumbongHermann🚀 Introducing Next-Concept Modeling (NCM): our new paradigm for fast, parallel, coarse-to-fine text generation. Compared to strong diffusion language models (DLMs) at similar parameter size, NCM achieves: ⚡ Up to 31× faster generation on downstream tasks like summarization and machine translation with comparable or better quality. 🏎️ 10–21× faster unconditional generation across TinyStories, LM1B, and OpenWebText, with comparable or better generative perplexity and using 3–5× fewer training tokens. 🧩 Better performance on reasoning tasks like Sudoku, outperforming both AR and diffusion baselines with fewer sampling steps. With @ishanskhare, @IddahMlauzi, @ekellbuch, @simran_s_arora, @Azaliamirh, and @HazyResearch. More details in the thread 🧵👇597711.7K
🚀 Introducing Next-Concept Modeling (NCM): our new paradigm for fast, parallel, coarse-to-fine text generation. Compared to strong diffusion language models (DLMs) at similar parameter size, NCM achieves: ⚡ Up to 31× faster generation on downstream tasks like summarization and machine translation with comparable or better quality. 🏎️ 10–21× faster unconditional generation across TinyStories, LM1B, and OpenWebText, with comparable or better generative perplexity and using 3–5× fewer training tokens. 🧩 Better performance on reasoning tasks like Sudoku, outperforming both AR and diffusion baselines with fewer sampling steps. With @ishanskhare, @IddahMlauzi, @ekellbuch, @simran_s_arora, @Azaliamirh, and @HazyResearch. More details in the thread 🧵👇622001.1K495K- Posts @StanfordAILab0062.4K
- Show more posts
