The SWE-Bench Illusion: When State-of-the-Art LLMs Remember Instead of Reason

• URL: http://arxiv.org/abs/2506.12286v3
• Date: Wed, 06 Aug 2025 11:23:11 GMT
• Title: The SWE-Bench Illusion: When State-of-the-Art LLMs Remember Instead of Reason
• Authors: Shanchao Liang, Spandan Garg, Roshanak Zilouchian Moghaddam,
• Abstract summary: We present empirical evidence that performance gains on SWE-Bench-Verified may be partially driven by memorization rather than genuine problem-solving.<n>We show that state-of-the-art models achieve up to 76% accuracy in identifying buggy file paths using only issue descriptions, without access to repository structure.<n>These findings raise concerns about the validity of existing results and underscore the need for more robust, contamination-resistant benchmarks.
• Score: 1.6249398255272318
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: As large language models (LLMs) become increasingly capable and widely adopted, benchmarks play a central role in assessing their practical utility. For example, SWE-Bench Verified has emerged as a critical benchmark for evaluating LLMs' software engineering abilities, particularly their aptitude for resolving real-world GitHub issues. Recent LLMs show impressive performance on SWE-Bench, leading to optimism about their capacity for complex coding tasks. However, current evaluation protocols may overstate these models' true capabilities. It is crucial to distinguish LLMs' generalizable problem-solving ability and other learned artifacts. In this work, we introduce two diagnostic tasks: file path identification from issue descriptions alone and ground truth function reproduction with only the current file context and issue description to probe models' underlying knowledge. We present empirical evidence that performance gains on SWE-Bench-Verified may be partially driven by memorization rather than genuine problem-solving. We show that state-of-the-art models achieve up to 76% accuracy in identifying buggy file paths using only issue descriptions, without access to repository structure. This performance is merely up to 53% on tasks from repositories not included in SWE-Bench, pointing to possible data contamination or memorization. Similar patterns are also observed for the function reproduction task, where the verbatim similarity is much higher on SWE-Bench Verified than on other similar coding benchmarks (up to 35% consecutive 5-gram accuracy on SWE-Bench Verified and Full, but only up to 18% for tasks in other benchmarks). These findings raise concerns about the validity of existing results and underscore the need for more robust, contamination-resistant benchmarks to reliably evaluate LLMs' coding abilities.

Related papers

• CompassVerifier: A Unified and Robust Verifier for LLMs Evaluation and Outcome Reward [50.97588334916863]
  We develop CompassVerifier, an accurate and robust lightweight verifier model for evaluation and outcome reward.<n>It demonstrates multi-domain competency spanning math, knowledge, and diverse reasoning tasks, with the capability to process various answer types.<n>We introduce VerifierBench benchmark comprising model outputs collected from multiple data sources, augmented through manual analysis of metaerror patterns to enhance CompassVerifier.
  arXiv  Detail & Related papers  (2025-08-05T17:55:24Z)
• SWE-Perf: Can Language Models Optimize Code Performance on Real-World Repositories? [32.67971774793393]
  SWE-Perf is the first benchmark designed to evaluate Large Language Models (LLMs) on code performance optimization tasks within authentic repository contexts.<n>SWE-Perf comprises 140 carefully curated instances, each derived from performance-improving pull requests from popular GitHub repositories.
  arXiv  Detail & Related papers  (2025-07-16T17:05:17Z)
• CodeJudgeBench: Benchmarking LLM-as-a-Judge for Coding Tasks [63.562924932512765]
  Large Language Models (LLMs) have advanced the state-of-the-art in various coding tasks.<n>LLMs can also serve as judges, assessing and comparing the quality of responses generated by other models.
  arXiv  Detail & Related papers  (2025-07-14T17:56:29Z)
• Verifying the Verifiers: Unveiling Pitfalls and Potentials in Fact Verifiers [59.168391398830515]
  We evaluate 12 pre-trained LLMs and one specialized fact-verifier, using a collection of examples from 14 fact-checking benchmarks.<n>We highlight the importance of addressing annotation errors and ambiguity in datasets.<n> frontier LLMs with few-shot in-context examples, often overlooked in previous works, achieve top-tier performance.
  arXiv  Detail & Related papers  (2025-06-16T10:32:10Z)
• OSS-Bench: Benchmark Generator for Coding LLMs [4.393587297483245]
  We introduce OSS-Bench, a benchmark generator that constructs large-scale, live evaluation tasks from real-world open-source software.<n> OSS-Bench replaces functions with LLM-generated code and evaluates them using three natural metrics: compilability, functional correctness, and memory safety.<n>Our results demonstrate that OSS-Bench mitigates overfitting by leveraging the evolving complexity of OSS.
  arXiv  Detail & Related papers  (2025-05-18T09:53:51Z)
• FEA-Bench: A Benchmark for Evaluating Repository-Level Code Generation for Feature Implementation [26.14778133391999]
  FEA-Bench is a benchmark designed to assess the ability of large language models to perform incremental development within code repositories.<n>We collect pull requests from 83 GitHub repositories and use rule-based and intent-based filtering to construct task instances focused on new feature development.
  arXiv  Detail & Related papers  (2025-03-09T16:11:57Z)
• Leveraging Online Olympiad-Level Math Problems for LLMs Training and Contamination-Resistant Evaluation [55.21013307734612]
  AoPS-Instruct is a dataset of more than 600,000 high-quality QA pairs.<n>LiveAoPSBench is an evolving evaluation set with timestamps, derived from the latest forum data.<n>Our work presents a scalable approach to creating and maintaining large-scale, high-quality datasets for advanced math reasoning.
  arXiv  Detail & Related papers  (2025-01-24T06:39:38Z)
• SWE-Fixer: Training Open-Source LLMs for Effective and Efficient GitHub Issue Resolution [56.9361004704428]
  Large Language Models (LLMs) have demonstrated remarkable proficiency across a variety of complex tasks.<n>SWE-Fixer is a novel open-source framework designed to effectively and efficiently resolve GitHub issues.<n>We assess our approach on the SWE-Bench Lite and Verified benchmarks, achieving competitive performance among open-source models.
  arXiv  Detail & Related papers  (2025-01-09T07:54:24Z)
• Are Your LLMs Capable of Stable Reasoning? [38.03049704515947]
  We introduce G-Pass@$k$, a novel evaluation metric that continuously assesses model performance across multiple sampling attempts.<n>We employ G-Pass@$k$ in conjunction with state-of-the-art large language models to provide comprehensive insights into their potential capabilities and operational consistency.
  arXiv  Detail & Related papers  (2024-12-17T18:12:47Z)
• A Real-World Benchmark for Evaluating Fine-Grained Issue Solving Capabilities of Large Language Models [11.087034068992653]
  FAUN-Eval is a benchmark specifically designed to evaluate the Fine-grAined issUe solviNg capabilities of LLMs.<n>It is constructed using a dataset curated from 30 well-known GitHub repositories.<n>We evaluate ten LLMs with FAUN-Eval, including four closed-source and six open-source models.
  arXiv  Detail & Related papers  (2024-11-27T03:25:44Z)
• BenTo: Benchmark Task Reduction with In-Context Transferability [32.561978389905434]
  This paper investigates how to efficiently reduce the tasks used to benchmark large language models (LLMs) We propose a practically efficient metric for estimating the transferability between two tasks via in-context learning (ICL)
  arXiv  Detail & Related papers  (2024-10-17T17:41:15Z)
• AutoDetect: Towards a Unified Framework for Automated Weakness Detection in Large Language Models [95.09157454599605]
  Large Language Models (LLMs) are becoming increasingly powerful, but they still exhibit significant but subtle weaknesses.<n>Traditional benchmarking approaches cannot thoroughly pinpoint specific model deficiencies.<n>We introduce a unified framework, AutoDetect, to automatically expose weaknesses in LLMs across various tasks.
  arXiv  Detail & Related papers  (2024-06-24T15:16:45Z)
• ML-Bench: Evaluating Large Language Models and Agents for Machine Learning Tasks on Repository-Level Code [76.84199699772903]
  ML-Bench is a benchmark rooted in real-world programming applications that leverage existing code repositories to perform tasks. To evaluate both Large Language Models (LLMs) and AI agents, two setups are employed: ML-LLM-Bench for assessing LLMs' text-to-code conversion within a predefined deployment environment, and ML-Agent-Bench for testing autonomous agents in an end-to-end task execution within a Linux sandbox environment.
  arXiv  Detail & Related papers  (2023-11-16T12:03:21Z)
• LLMs as Factual Reasoners: Insights from Existing Benchmarks and Beyond [135.8013388183257]
  We propose a new protocol for inconsistency detection benchmark creation and implement it in a 10-domain benchmark called SummEdits. Most LLMs struggle on SummEdits, with performance close to random chance. The best-performing model, GPT-4, is still 8% below estimated human performance.
  arXiv  Detail & Related papers  (2023-05-23T21:50:06Z)

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.