Teaching LLMs to reason like Bayesians
Google researchers demonstrate how Bayesian teaching through supervised fine-tuning enables LLMs to approximate optimal probabilistic reasoning and generalize to new domains.
Teaching LLMs to reason like Bayesians Skip to main content Teaching LLMs to reason like Bayesians March 4, 2026 Sjoerd van Steenkiste and Tal Linzen, Research Scientists, Google Research We teach LLMs to reason in a Bayesian manner by training them to mimic the predictions of an optimal Bayesian model. Quick links Paper Share Copy link × AI systems based on large language models (LLMs) are increasingly used as agents that interact with users and the world. To do this successfully, LLMs need to construct internal representations of the world and estimate the probability that each of these repr
saved by
related reading
- On-Policy Distillation - Thinking Machines Labthinkingmachines.ai
- Bayesian Neural Networkscs.toronto.edu
- Modifying LLM Beliefs with Synthetic Document Finetuningalignment.anthropic.com
- How does in-context learning work? A framework for understanding the differences from traditional supervised learning | SAIL Blogai.stanford.edu
- [2305.18290] Direct Preference Optimization: Your Language Model is Secretly a Reward Modelarxiv.org
- Base Models Know How to Reason, Thinking Models Learn Whenarxiv.org
- State of RL for reasoning LLMs | A. Weersaweers.de
- You can’t imitation-learn how to continual-learn — LessWronglesswrong.com
- Large Language Models Must Be Taught to Know What They Don't Knowarxiv.org
- The State of Reinforcement Learning for LLM Reasoningmagazine.sebastianraschka.com
- [2511.08579] Training Language Models to Explain Their Own Computationsarxiv.org
- [2510.13786] The Art of Scaling Reinforcement Learning Compute for LLMsarxiv.org