On Memorization of Large Language Models in Logical Reasoning

• URL: http://arxiv.org/abs/2410.23123v1
• Date: Wed, 30 Oct 2024 15:31:54 GMT
• Title: On Memorization of Large Language Models in Logical Reasoning
• Authors: Chulin Xie, Yangsibo Huang, Chiyuan Zhang, Da Yu, Xinyun Chen, Bill Yuchen Lin, Bo Li, Badih Ghazi, Ravi Kumar,
• Abstract summary: Large language models (LLMs) achieve good performance on challenging reasoning benchmarks, yet could also make basic reasoning mistakes. One hypothesis is that the increasingly high and nearly saturated performance could be due to the memorization of similar problems. We show that fine-tuning leads to heavy memorization, but it also consistently improves generalization performance.
• Score: 70.94164038947078
• License:
• Abstract: Large language models (LLMs) achieve good performance on challenging reasoning benchmarks, yet could also make basic reasoning mistakes. This contrasting behavior is puzzling when it comes to understanding the mechanisms behind LLMs' reasoning capabilities. One hypothesis is that the increasingly high and nearly saturated performance on common reasoning benchmarks could be due to the memorization of similar problems. In this paper, we systematically investigate this hypothesis with a quantitative measurement of memorization in reasoning tasks, using a dynamically generated logical reasoning benchmark based on Knights and Knaves (K&K) puzzles. We found that LLMs could interpolate the training puzzles (achieving near-perfect accuracy) after fine-tuning, yet fail when those puzzles are slightly perturbed, suggesting that the models heavily rely on memorization to solve those training puzzles. On the other hand, we show that while fine-tuning leads to heavy memorization, it also consistently improves generalization performance. In-depth analyses with perturbation tests, cross difficulty-level transferability, probing model internals, and fine-tuning with wrong answers suggest that the LLMs learn to reason on K&K puzzles despite training data memorization. This phenomenon indicates that LLMs exhibit a complex interplay between memorization and genuine reasoning abilities. Finally, our analysis with per-sample memorization score sheds light on how LLMs switch between reasoning and memorization in solving logical puzzles. Our code and data are available at https://memkklogic.github.io.

Related papers

• Do Large Language Models Truly Grasp Mathematics? An Empirical Exploration From Cognitive Psychology [13.964263002704582]
  We show that, even with the use of Chains of Thought prompts, mainstream LLMs have a high error rate when solving modified CRT problems. Specifically, the average accuracy rate dropped by up to 50% compared to the original questions. This finding challenges the belief that LLMs have genuine mathematical reasoning abilities comparable to humans.
  arXiv  Detail & Related papers  (2024-10-19T05:01:56Z)
• Not All LLM Reasoners Are Created Equal [58.236453890457476]
  We study the depth of grade-school math problem-solving capabilities of LLMs. We evaluate their performance on pairs of existing math word problems together.
  arXiv  Detail & Related papers  (2024-10-02T17:01:10Z)
• Step-by-Step Reasoning to Solve Grid Puzzles: Where do LLMs Falter? [36.14795256060537]
  We develop GridPuzzle, an evaluation dataset comprising 274 grid-based puzzles with different complexities. Second, we propose a new error taxonomy derived from manual analysis of reasoning chains from LLMs including GPT-4, Claude-3, Gemini, Mistral, and Llama-2. Third, we develop an LLM-based framework for large-scale subjective evaluation (i.e., identifying errors) and an objective metric, PuzzleEval, to evaluate the correctness of reasoning chains.
  arXiv  Detail & Related papers  (2024-07-20T07:43:07Z)
• Improving Complex Reasoning over Knowledge Graph with Logic-Aware Curriculum Tuning [89.89857766491475]
  We propose a complex reasoning schema over KG upon large language models (LLMs) We augment the arbitrary first-order logical queries via binary tree decomposition to stimulate the reasoning capability of LLMs. Experiments across widely used datasets demonstrate that LACT has substantial improvements(brings an average +5.5% MRR score) over advanced methods.
  arXiv  Detail & Related papers  (2024-05-02T18:12:08Z)
• LogicBench: Towards Systematic Evaluation of Logical Reasoning Ability of Large Language Models [52.03659714625452]
  Recently developed large language models (LLMs) have been shown to perform remarkably well on a wide range of language understanding tasks. But, can they really "reason" over the natural language? This question has been receiving significant research attention and many reasoning skills such as commonsense, numerical, and qualitative have been studied.
  arXiv  Detail & Related papers  (2024-04-23T21:08:49Z)
• Logic Query of Thoughts: Guiding Large Language Models to Answer Complex Logic Queries with Knowledge Graphs [102.37496443389203]
  'Logic-Query-of-Thoughts' (LGOT) is the first of its kind to combine knowledge graph reasoning and large language models. Our experimental findings demonstrate substantial performance enhancements, with up to 20% improvement over ChatGPT.
  arXiv  Detail & Related papers  (2024-03-17T17:01:45Z)
• GSM-Plus: A Comprehensive Benchmark for Evaluating the Robustness of LLMs as Mathematical Problem Solvers [68.77382332826167]
  Large language models (LLMs) have achieved impressive performance across various mathematical reasoning benchmarks. One essential and frequently occurring evidence is that when the math questions are slightly changed, LLMs can behave incorrectly. This motivates us to evaluate the robustness of LLMs' math reasoning capability by testing a wide range of question variations.
  arXiv  Detail & Related papers  (2024-02-29T15:26:14Z)
• Do Large Language Models Understand Logic or Just Mimick Context? [14.081178100662163]
  This paper investigates the reasoning capabilities of large language models (LLMs) on two logical reasoning datasets. It is found that LLMs do not truly understand logical rules; rather, in-context learning has simply enhanced the likelihood of these models arriving at the correct answers.
  arXiv  Detail & Related papers  (2024-02-19T12:12:35Z)
• Improving Large Language Models in Event Relation Logical Prediction [33.88499005859982]
  Event relation extraction is a challenging task that demands thorough semantic understanding and rigorous logical reasoning. In this paper, we conduct an in-depth investigation to systematically explore the capability of LLMs in understanding and applying event relation logic. Our study reveals that LLMs are not logically consistent reasoners, which results in their suboptimal performance on tasks that need rigorous reasoning.
  arXiv  Detail & Related papers  (2023-10-13T14:53:06Z)

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.