Reasoning

Getting a model to show its work, and checking the work.

6 entries

  1. 2026ICML 2026PublishedCurated

    Beyond Trajectory-Level Attribution: Graph-Based Credit Assignment for Agentic Reinforcement Learning

    Cheng et al.

    Group-based reinforcement learning methods credit whole trajectories by final outcome, which makes it hard to see the contribution of individual steps — including valuable steps buried inside failed trajectories. The authors propose Graph-based Group Policy Optimization (GraphGPO), which aggregates all rollout trajectories into a unified state-transition graph, estimates the distance from each state to the task goal using global information encoded in the graph, and assigns credit to each edge by how much the transition reduces that distance. They report improved training efficiency and state-of-the-art performance across a range of benchmarks.

    ReasoningAdaptation
    arXiv
  2. 2023NeurIPS 2023PublishedCurated

    Tree of Thoughts: Deliberate Problem Solving with Large Language Models

    Yao et al.

    Generalises chain-of-thought into a search tree. The model proposes multiple partial solutions, evaluates them against each other, and can back out of a branch that is not working, rather than committing to one linear line of reasoning.

    ReasoningAgents & Tool Use
    arXiv
  3. 2023NeurIPS 2023PublishedCurated

    Reflexion: Language Agents with Verbal Reinforcement Learning

    Shinn et al.

    Instead of updating weights, the agent writes down what went wrong after a failed attempt and carries that reflection into the next attempt. The reflection is stored in natural language and used as additional context, so no gradient step is required.

    Agents & Tool UseReasoning
    arXiv
  4. 2022ICLR 2023PublishedCurated

    ReAct: Synergizing Reasoning and Acting in Language Models

    Yao et al.

    Interleaves verbal reasoning traces with tool actions, so the model thinks about what to do next, does it, then reasons about what came back. The reasoning trace also becomes an auditable record of why the agent took each step.

    Agents & Tool UseReasoning
    arXiv
  5. 2022ICLR 2023PublishedCurated

    Self-Consistency Improves Chain of Thought Reasoning in Language Models

    Wang et al.

    Samples several independent reasoning paths for the same question and takes the most common answer, instead of trusting a single greedy chain. The intuition is that a correct derivation can be reached many ways while a wrong one is usually reached once.

    Reasoning
    arXiv
  6. 2022NeurIPS 2022PublishedCurated

    Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

    Wei et al.

    Shows that asking a model to write out intermediate reasoning steps before answering substantially improves multi-step arithmetic, commonsense and symbolic reasoning, and that the behaviour appears without any fine-tuning at sufficient model scale.

    Reasoning
    arXiv

All research