Language-model search improves by skipping repeated reasoning errors
CORE: Conflict-Oriented Reasoning Elimination for Verifiable Language-Model Search
Computation and Language
Summary
When language models try to solve problems step-by-step, they often restart from the last step when they fail, missing earlier mistakes that caused the issue. The authors introduce CORE, a method that identifies the key earlier decisions that led to failure and jumps back directly to fix those, avoiding wasted effort. This approach also remembers these problem areas to prevent repeated mistakes. Experiments show that CORE makes the search process more efficient and increases success rates compared to prior methods.
What this means in practice
- •For language model engineers: Improve reasoning tasks in language models by integrating CORE to reduce redundant verification calls and increase success rates.
- •For ai platform developers: Develop more efficient language-model-powered tools by using CORE to guide search away from earlier erroneous steps, saving time and computation.
Authors
Siyu Song, Rui Xu, Jia Lin, Kai Liu, Weifang Wang
Abstract
Test-time reasoning systems often respond to failure by restarting or revising the latest step, even when an earlier decision caused the error. We introduce CORE, a search controller that requests a certified conflict core from a verifier, backjumps to the latest decision in that core, and caches the conflict to avoid repeating it. Under sound verification, finite branching and depth, and exhaustive proposals, the uncapped search is complete and never prunes a valid solution. On 2,000 planted graph-coloring instances with matched proposals and an exact verifier, CORE reduces median verifier calls by 39.8% at 30 variables and 35.0% at 36 variables relative to chronological repair; caching further improves on backjumping alone. Across five reasoning tasks, CORE achieves 75.9% mean success with Qwen2.5-7B-Instruct and 84.2% with Qwen3-8B, compared with 72.5% and 81.8% for Tree of Thoughts. It also uses fewer verifier calls and generated tokens on both backbones. These results show the value of using certified failure explanations to direct language-model search.