Inducing Reasoning Primitives from Agent Traces

Zhihan Lei; Jiarui Yan; Joshua Momo; William W. Cohen

cs.AIJun 2, 2026

Inducing Reasoning Primitives from Agent Traces

Authors: Zhihan Lei, Jiarui Yan, Joshua Momo, William W. Cohen

Organizations: Carnegie Mellon University

Abstract

ReAct-style LLM agents often rediscover the same reasoning routines across problems, yet leave those routines trapped in transient scratchpads. We introduce Reasoning Primitive Induction, a single-pass method that mines successful ReAct traces, clusters recurrent reasoning moves, and converts the most frequent moves into a compact library of typed pseudo-tools. Each pseudo-tool is specified by a natural-language docstring interpreted by an LLM at invocation time, and a standard ReAct loop composes these primitives at test time. The central result is that induced libraries outperform the very agent that generated their traces: by +44pp on RuleArena NBA (30 -> 74), +30pp on MuSR team allocation (38 -> 68), and +22pp on NatPlan meeting planning (7 -> 29). Across five comparable subtasks spanning narrative deduction, rule application, and constraint-satisfaction planning, a single fixed configuration improves over zero-shot Chain-of-Thought on every subtask, matches or surpasses expert-authored decompositions, and outperforms AWM at lower average inference cost.

Explore similar work

RAG over Thinking Traces Can Improve Reasoning Tasks
May 5, 2026Negar Arabzadeh, Wenjie Ma, Sewon Min +1Reasoning Traces
ReasonOps: Operator Segmentation for LLM Reasoning Traces
May 28, 2026Daniel Lee, Owen Queen, James ZouReasoning Traces Large Language Model Reasoning
PetriBench: Benchmarking LLM Reasoning over Dynamic State Spaces
Sep 17, 2026Pyrros Koussios, Benjamin Jäger, John Hua Yao +3Reasoning Benchmark Large Language Model Reasoning

May 5, 2026 · cs.IRJ/K move · Enter open · S save

RAG over Thinking Traces Can Improve Reasoning Tasks

Negar Arabzadeh, Wenjie Ma, Sewon Min, Matei Zaharia

University of California, Berkeley

Retrieval-augmented generation (RAG) has proven effective for knowledge-intensive tasks, but is widely believed to offer limited benefit for reasoning-intensive problems such as math and code generation. We challenge this assumption by showing that the limitation lies not in RAG itself, but in the choice of corpus. Instead of retrieving documents, we propose retrieving thinking traces, i.e., intermediate thinking trajectories generated during problem solving attempts. We show that thinking traces are already a strong retrieval source, and further introduce T3, an offline method that transforms them into structured, retrieval-friendly representations, to improve usability. Using these traces as a corpus, a simple retrieve-then-generate pipeline consistently improves reasoning performance across strong models and benchmarks such as AIME 2025--2026, LiveCodeBench, and GPQA-Diamond, outperforming both non-RAG baselines and retrieval over standard web corpora. For instance, on AIME 2025-2026, RAG with traces generated by Gemini-2-thinking achieves relative gains of +56.3%, +8.6%, and +7.6% for Gemini-2.5-Flash, GPT-OSS-120B, and GPT-5, respectively, even though these are more recent models. Overall, our results suggest that thinking traces are an effective retrieval corpus for reasoning tasks, and transforming them into structured, compact, or diagnostic representations unlocks even stronger gains. Code available at https://github.com/Narabzad/t3.

Reasoning Traces

Open paper Ask about it PDF

Abstract

Explore similar work

RAG over Thinking Traces Can Improve Reasoning Tasks

ReasonOps: Operator Segmentation for LLM Reasoning Traces

PetriBench: Benchmarking LLM Reasoning over Dynamic State Spaces