Constitutional AI showed LMs can learn to follow constitutions by labeling their own outputs. But why can't we just tell a base model the principles of desired behavior and rely on it to act appropriately?
Introducing SAMI: Self-Supervised Alignment with Mutual Information!
Excited to share OffTheRails: A moral reasoning benchmark beyond trolley problems!
We present a simple prompting pipeline for generating moral reasoning evaluations with language models using causal templates 🔵→🟠
Language models struggle to search, not due to an architecture problem, but a data one! They rarely see how to search or backtrack. We show how LLMs can be taught to search by representing the process of search in language as a flattened string, a stream of search (SoS)!
Multi-turn interactive RL should be a bigger focus. Current methods are not well-suited for this - i.e. PPO can't train with user in the loop generally and offline Q-learning still does not work at scale. It's interesting to see more work in that direction.
Multi-turn interactive RL should be a bigger focus. Current methods are not well-suited for this - i.e. PPO can't train with user in the loop generally and offline Q-learning still does not work at scale. It's interesting to see more work in that direction.
705 Followers 691 FollowingPostdoc at @Stanford, @StanfordCISAC, Stanford Center for AI Safety, SERI. | Focusing on interpretable, safe, and ethical AI/LLM decision-making. Find me on 🦋
3K Followers 6K FollowingLLM for code and reasoning. PhD student at Cornell. Previously Student Researcher at @google. Previously intern at @theteamatx.
2K Followers 837 FollowingAI Researcher @CapitalOne AIF. Ex @TechAtBloomberg @BigScienceW @SFResearch @hkust. Working on multilingual and LLM #NLProc. Building @GrassrootsSci
10 Followers 103 FollowingJe voudrais que quelqu'un m'attende quelque part, laissez le temps ressentir la température du mot, et laissez retentir longtemps...
165 Followers 384 FollowingResearch scientist at Amazon. Interested in language models and responsible AI. Studied at @Mila_Quebec during my Ph.D. and interned at Microsoft Research.
476 Followers 3K FollowingPhD student @PurdueECE, researching deep learning optimization theory and intrinsic interpretability. I love open source. @jinen:https://t.co/W0XuIlDIe9
146 Followers 1K FollowingComputational modeling of human learning: cognitive development, language acquisition, social learning, causal learning... Brown PhD student with @banhpad
66K Followers 1K FollowingRunning for Congress to represent San Francisco. No corporate or lobbyist money. Past: CoS to AOC, Dir. Tech @ Bernie, founding engineer @stripe.
600 Followers 270 FollowingFounder at https://t.co/RxBIw3OY9x | AI with first principles in engineering, biology, robotics, autonomous systems, and LLMs | @USC alum, @JohnsHopkins dropout
705 Followers 691 FollowingPostdoc at @Stanford, @StanfordCISAC, Stanford Center for AI Safety, SERI. | Focusing on interpretable, safe, and ethical AI/LLM decision-making. Find me on 🦋
617K Followers 981 FollowingDemocratic Nominee for Mayor of NYC. Assemblymember. Running to freeze the rent, make buses fast + free, and deliver universal childcare. Democratic Socialist.
3K Followers 6K FollowingLLM for code and reasoning. PhD student at Cornell. Previously Student Researcher at @google. Previously intern at @theteamatx.
20K Followers 452 Followingphysics of language models @ Meta (FAIR, not GenAI)
🎓:Tsinghua Physics — MIT CSAIL — Princeton/IAS
🏅:IOI x 2 — ACM-ICPC — USACO — Codejam — math MCM