Subtle Errors Matter: Preference Learning via Error-injected Self-editing

• URL: http://arxiv.org/abs/2410.06638v3
• Date: Mon, 03 Mar 2025 07:09:42 GMT
• Title: Subtle Errors Matter: Preference Learning via Error-injected Self-editing
• Authors: Kaishuai Xu, Tiezheng Yu, Wenjun Hou, Yi Cheng, Chak Tou Leong, Liangyou Li, Xin Jiang, Lifeng Shang, Qun Liu, Wenjie Li,
• Abstract summary: We propose a novel preference learning framework called eRror-Injected Self-Editing (RISE)<n>RISE injects predefined subtle errors into pivotal tokens in reasoning or steps to construct hard pairs for error mitigation.<n>Experiments validate the effectiveness of RISE, with preference learning on Qwen2-7B-Instruct yielding notable improvements of 3.0% on GSM8K and 7.9% on MATH with only 4.5K training samples.
• Score: 59.405145971637204
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Models (LLMs) have exhibited strong mathematical reasoning prowess, tackling tasks ranging from basic arithmetic to advanced competition-level problems. However, frequently occurring subtle yet critical errors, such as miscalculations or incorrect substitutions, limit the LLMs' full potential. Existing studies to improve mathematical ability typically involve applying preference learning to step-wise solution pairs. Although these methods leverage samples of varying granularity to mitigate reasoning errors, they overlook critical subtle errors. In this work, we propose a novel preference learning framework called eRror-Injected Self-Editing (RISE), which injects predefined subtle errors into pivotal tokens in reasoning or computation steps to construct hard pairs for error mitigation. In detail, RISE uses the LLM itself to edit a small number of tokens in the solution, injecting designed subtle errors. Then, pairs composed of self-edited solutions and their corresponding correct ones, along with pairs of correct and incorrect solutions obtained through sampling, are used together for subtle error-aware DPO training. Compared with other preference learning methods, RISE further refines the training objective without requiring fine-grained sampling or preference annotation. Extensive experiments validate the effectiveness of RISE, with preference learning on Qwen2-7B-Instruct yielding notable improvements of 3.0% on GSM8K and 7.9% on MATH with only 4.5K training samples. Moreover, the effect of error mitigation extends from mathematical reasoning to logical reasoning and code generation.

Related papers

• ZeroED: Hybrid Zero-shot Error Detection through Large Language Model Reasoning [45.352592886478774]
  We propose ZeroED, a novel hybrid zero-shot error detection framework. ZeroED operates in four steps, i.e., feature representation, error labeling, training data construction, and detector training. Experiments show ZeroED substantially outperforms state-of-the-art methods by a maximum 30% improvement in F1 score and up to 90% token cost reduction.
  arXiv  Detail & Related papers  (2025-04-06T10:28:41Z)
• LEMMA: Learning from Errors for MatheMatical Advancement in LLMs [33.571479131705075]
  We introduce Learning from Errors for Mathematical Advancement (LEMMA) to enhance large language models' reasoning ability. LEMMA constructs data consisting of an incorrect solution with an erroneous step and a reflection connection to a correct solution for fine-tuning. Experimental results demonstrate that LEMMA achieves significant performance improvements over other strong baselines.
  arXiv  Detail & Related papers  (2025-03-21T17:59:10Z)
• IterPref: Focal Preference Learning for Code Generation via Iterative Debugging [28.020886216989872]
  We propose IterPref, a new preference alignment framework for Code LLMs. IterPref explicitly locates error regions and aligns the corresponding tokens via a tailored DPO algorithm. IterPref achieves significant performance gains in code generation and improves on challenging tasks like BigCodeBench.
  arXiv  Detail & Related papers  (2025-03-04T16:56:34Z)
• Error Classification of Large Language Models on Math Word Problems: A Dynamically Adaptive Framework [64.83955753606443]
  Math Word Problems serve as a crucial benchmark for evaluating Large Language Models' reasoning abilities. Current error classification methods rely on static and predefined categories. We introduce MWPES-300K, a comprehensive dataset containing 304,865 error samples.
  arXiv  Detail & Related papers  (2025-01-26T16:17:57Z)
• Probably Approximately Precision and Recall Learning [62.912015491907994]
  Precision and Recall are foundational metrics in machine learning. One-sided feedback--where only positive examples are observed during training--is inherent in many practical problems. We introduce a PAC learning framework where each hypothesis is represented by a graph, with edges indicating positive interactions.
  arXiv  Detail & Related papers  (2024-11-20T04:21:07Z)
• Improving LLM Reasoning through Scaling Inference Computation with Collaborative Verification [52.095460362197336]
  Large language models (LLMs) struggle with consistent and accurate reasoning. LLMs are trained primarily on correct solutions, reducing their ability to detect and learn from errors. We propose a novel collaborative method integrating Chain-of-Thought (CoT) and Program-of-Thought (PoT) solutions for verification.
  arXiv  Detail & Related papers  (2024-10-05T05:21:48Z)
• S$^3$c-Math: Spontaneous Step-level Self-correction Makes Large Language Models Better Mathematical Reasoners [23.713779973116733]
  Self-correction is a method that can stimulate the potential reasoning abilities of large language models (LLMs) We propose S$3$c-Math, which are able to perform Spontaneous Step-level Self-correction for Mathematical reasoning.
  arXiv  Detail & Related papers  (2024-09-03T01:40:21Z)
• Physics of Language Models: Part 2.2, How to Learn From Mistakes on Grade-School Math Problems [47.753284211200665]
  We focus on understanding the usefulness of incorporating "error-correction" data directly into the pretraining stage. This data consists of erroneous solution steps immediately followed by their corrections. We show promising results: this type of pretrain data can help language models achieve higher reasoning accuracy.
  arXiv  Detail & Related papers  (2024-08-29T06:49:20Z)
• Learning to Correct for QA Reasoning with Black-box LLMs [37.13135300208977]
  We propose CoBB (Correct for improving QA reasoning of Black-Box LLMs) as an open challenge in machine learning. It uses a trained adaptation model to perform a seq2seq mapping from the often-imperfect reasonings of the original black-box LLM to the correct or improved reasonings. Our experimental results demonstrate that CoBB significantly improves reasoning accuracy across various QA benchmarks.
  arXiv  Detail & Related papers  (2024-06-26T18:57:32Z)
• LLM Critics Help Catch Bugs in Mathematics: Towards a Better Mathematical Verifier with Natural Language Feedback [71.95402654982095]
  We propose Math-Minos, a natural language feedback-enhanced verifier. Our experiments reveal that a small set of natural language feedback can significantly boost the performance of the verifier.
  arXiv  Detail & Related papers  (2024-06-20T06:42:27Z)
• Evaluating Mathematical Reasoning of Large Language Models: A Focus on Error Identification and Correction [35.01097297297534]
  Existing evaluations of Large Language Models (LLMs) focus on problem-solving from the examinee perspective. We define four evaluation tasks for error identification and correction along with a new dataset with annotated error types and steps. Our principal findings indicate that GPT-4 outperforms all models, while open-source model LLaMA-2-7B demonstrates comparable abilities to closed-source models GPT-3.5 and Gemini Pro.
  arXiv  Detail & Related papers  (2024-06-02T14:16:24Z)
• Small Language Models Need Strong Verifiers to Self-Correct Reasoning [69.94251699982388]
  Self-correction has emerged as a promising solution to boost the reasoning performance of large language models (LLMs) This work explores whether small (= 13B) language models (LMs) have the ability of self-correction on reasoning tasks with minimal inputs from stronger LMs.
  arXiv  Detail & Related papers  (2024-04-26T03:41:28Z)
• Achieving >97% on GSM8K: Deeply Understanding the Problems Makes LLMs Better Solvers for Math Word Problems [50.76385564061713]
  Chain-of-Thought (CoT) prompting has enhanced the performance of Large Language Models (LLMs) across various reasoning tasks. CoT usually suffers from three pitfalls: semantic misunderstanding errors, calculation errors, and step-missing errors. We propose Deeply Understanding the Problems (DUP) to improve the LLMs' math problem-solving ability by addressing semantic misunderstanding errors.
  arXiv  Detail & Related papers  (2024-04-23T12:16:05Z)
• Understanding and Mitigating Classification Errors Through Interpretable Token Patterns [58.91023283103762]
  Characterizing errors in easily interpretable terms gives insight into whether a classifier is prone to making systematic errors. We propose to discover those patterns of tokens that distinguish correct and erroneous predictions. We show that our method, Premise, performs well in practice.
  arXiv  Detail & Related papers  (2023-11-18T00:24:26Z)
• Learning From Mistakes Makes LLM Better Reasoner [106.48571828587728]
  Large language models (LLMs) recently exhibited remarkable reasoning capabilities on solving math problems. This work explores whether LLMs can LEarn from MistAkes (LEMA), akin to the human learning process.
  arXiv  Detail & Related papers  (2023-10-31T17:52:22Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.