SUPER: Evaluating Agents on Setting Up and Executing Tasks from Research Repositories

• URL: http://arxiv.org/abs/2409.07440v1
• Date: Wed, 11 Sep 2024 17:37:48 GMT
• Title: SUPER: Evaluating Agents on Setting Up and Executing Tasks from Research Repositories
• Authors: Ben Bogin, Kejuan Yang, Shashank Gupta, Kyle Richardson, Erin Bransom, Peter Clark, Ashish Sabharwal, Tushar Khot,
• Abstract summary: Super aims to capture the realistic challenges faced by researchers working with Machine Learning (ML) and Natural Language Processing (NLP) research repositories. Our benchmark comprises three distinct problem sets: 45 end-to-end problems with annotated expert solutions, 152 sub problems derived from the expert set that focus on specific challenges, and 602 automatically generated problems for larger-scale development. We show that state-of-the-art approaches struggle to solve these problems with the best model (GPT-4o) solving only 16.3% of the end-to-end set, and 46.1% of the scenarios.
• Score: 55.161075901665946
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Given that Large Language Models (LLMs) have made significant progress in writing code, can they now be used to autonomously reproduce results from research repositories? Such a capability would be a boon to the research community, helping researchers validate, understand, and extend prior work. To advance towards this goal, we introduce SUPER, the first benchmark designed to evaluate the capability of LLMs in setting up and executing tasks from research repositories. SUPERaims to capture the realistic challenges faced by researchers working with Machine Learning (ML) and Natural Language Processing (NLP) research repositories. Our benchmark comprises three distinct problem sets: 45 end-to-end problems with annotated expert solutions, 152 sub problems derived from the expert set that focus on specific challenges (e.g., configuring a trainer), and 602 automatically generated problems for larger-scale development. We introduce various evaluation measures to assess both task success and progress, utilizing gold solutions when available or approximations otherwise. We show that state-of-the-art approaches struggle to solve these problems with the best model (GPT-4o) solving only 16.3% of the end-to-end set, and 46.1% of the scenarios. This illustrates the challenge of this task, and suggests that SUPER can serve as a valuable resource for the community to make and measure progress.

Related papers

• BeyondSWE: Can Current Code Agent Survive Beyond Single-Repo Bug Fixing? [61.247730037229815]
  We introduce BeyondSWE, a comprehensive benchmark that broadens existing evaluations along two axes - resolution scope and knowledge scope.<n>To investigate the role of external knowledge, we develop SearchSWE, a framework that integrates deep search with coding abilities.<n>This work offers both a realistic, challenging evaluation benchmark and a flexible framework to advance research toward more capable code agents.
  arXiv  Detail & Related papers  (2026-03-03T17:52:01Z)
• Deep Research: A Systematic Survey [118.82795024422722]
  Deep Research (DR) aims to combine the reasoning capabilities of large language models with external tools, such as search engines.<n>This survey presents a comprehensive and systematic overview of deep research systems.
  arXiv  Detail & Related papers  (2025-11-24T15:28:28Z)
• AInstein: Assessing the Feasibility of AI-Generated Approaches to Research Problems [28.38783951577184]
  AInstein is a framework for testing whether large language models can generate valid solutions to AI research problems.<n>We evaluate AInstein on 1,214 ICLR papers stratified by acceptance tier.
  arXiv  Detail & Related papers  (2025-10-06T22:50:41Z)
• IMProofBench: Benchmarking AI on Research-Level Mathematical Proof Generation [4.991157581428135]
  IMProofBench is a private benchmark consisting of 39 peer-reviewed problems developed by expert mathematicians.<n>Each problem requires a detailed proof and is paired with subproblems that have final answers.<n>Unlike prior benchmarks, the evaluation setup simulates a realistic research environment.
  arXiv  Detail & Related papers  (2025-09-30T10:50:37Z)
• Eigen-1: Adaptive Multi-Agent Refinement with Monitor-Based RAG for Scientific Reasoning [53.45095336430027]
  We develop a unified framework that combines implicit retrieval and structured collaboration.<n>On Humanity's Last Exam (HLE) Bio/Chem Gold, our framework achieves 48.3% accuracy.<n>Results on SuperGPQA and TRQA confirm robustness across domains.
  arXiv  Detail & Related papers  (2025-09-25T14:05:55Z)
• Exploring Autonomous Agents: A Closer Look at Why They Fail When Completing Tasks [8.218266805768687]
  We present a benchmark of 34 representative programmable tasks designed to rigorously assess autonomous agents.<n>We evaluate three popular open-source agent frameworks combined with two LLM backbones, observing a task completion rate of approximately 50%.<n>We develop a three-tier taxonomy of failure causes aligned with task phases, highlighting planning errors, task execution issues, and incorrect response generation.
  arXiv  Detail & Related papers  (2025-08-18T17:55:22Z)
• Trae Agent: An LLM-based Agent for Software Engineering with Test-time Scaling [18.390443362388623]
  Trae Agent is the first agent-based ensemble reasoning approach for repository-level issue resolution.<n>We conduct experiments using three leading large language models (LLMs) on the widely-adopted SWE-bench benchmark.<n>Trae Agent consistently achieves superior performance, with an average improvement of 10.22% over all baselines in terms of Pass@1.
  arXiv  Detail & Related papers  (2025-07-31T09:37:22Z)
• Which Agent Causes Task Failures and When? On Automated Failure Attribution of LLM Multi-Agent Systems [50.29939179830491]
  Failure attribution in LLM multi-agent systems remains underexplored and labor-intensive.<n>We develop and evaluate three automated failure attribution methods, summarizing their corresponding pros and cons.<n>The best method achieves 53.5% accuracy in identifying failure-responsible agents but only 14.2% in pinpointing failure steps.
  arXiv  Detail & Related papers  (2025-04-30T23:09:44Z)
• MLRC-Bench: Can Language Agents Solve Machine Learning Research Challenges? [64.62421656031128]
  MLRC-Bench is a benchmark designed to quantify how effectively language agents can tackle challenging Machine Learning (ML) Research Competitions. Unlike prior work, MLRC-Bench measures the key steps of proposing and implementing novel research methods. Even the best-performing tested agent closes only 9.3% of the gap between baseline and top human participant scores.
  arXiv  Detail & Related papers  (2025-04-13T19:35:43Z)
• Why Do Multi-Agent LLM Systems Fail? [91.39266556855513]
  We present MAST (Multi-Agent System Failure taxonomy), the first empirically grounded taxonomy designed to understand MAS failures. We analyze seven popular MAS frameworks across over 200 tasks, involving six expert human annotators. We identify 14 unique failure modes, organized into 3 overarching categories, (i) specification issues, (ii) inter-agent misalignment, and (iii) task verification.
  arXiv  Detail & Related papers  (2025-03-17T19:04:38Z)
• How to Get Your LLM to Generate Challenging Problems for Evaluation [33.625052642068624]
  CHASE is a unified framework to synthetically generate challenging problems using Large Language Models. We implement CHASE to create evaluation benchmarks across three diverse domains. The performance of state-of-the-art LLMs on these synthetic benchmarks lies in the range of 40-60% accuracy.
  arXiv  Detail & Related papers  (2025-02-20T16:09:55Z)
• MMAU: A Holistic Benchmark of Agent Capabilities Across Diverse Domains [54.117238759317004]
  Massive Multitask Agent Understanding (MMAU) benchmark features comprehensive offline tasks that eliminate the need for complex environment setups. It evaluates models across five domains, including Tool-use, Directed Acyclic Graph (DAG) QA, Data Science and Machine Learning coding, Contest-level programming and Mathematics. With a total of 20 meticulously designed tasks encompassing over 3K distinct prompts, MMAU provides a comprehensive framework for evaluating the strengths and limitations of LLM agents.
  arXiv  Detail & Related papers  (2024-07-18T00:58:41Z)
• Learning Task Decomposition to Assist Humans in Competitive Programming [90.4846613669734]
  We introduce a novel objective for learning task decomposition, termed value (AssistV) We collect a dataset of human repair experiences on different decomposed solutions. Under 177 hours of human study, our method enables non-experts to solve 33.3% more problems, speeds them up by 3.3x, and empowers them to match unassisted experts.
  arXiv  Detail & Related papers  (2024-06-07T03:27:51Z)
• PECC: Problem Extraction and Coding Challenges [3.287942619833188]
  We introduce PECC, a novel benchmark derived from Advent Of Code (AoC) challenges and Project Euler. Unlike conventional benchmarks, PECC requires LLMs to interpret narrative-embedded problems, extract requirements, and generate code. Results show varying model performance between narrative and neutral problems, with specific challenges in the Euler math-based subset.
  arXiv  Detail & Related papers  (2024-04-29T15:02:14Z)
• Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision [98.97575836717931]
  Current AI alignment methodologies rely on human-provided demonstrations or judgments. This raises a challenging research question: How can we keep improving the systems when their capabilities have surpassed the levels of humans?
  arXiv  Detail & Related papers  (2024-03-14T15:12:38Z)
• Effectiveness Assessment of Recent Large Vision-Language Models [78.69439393646554]
  This paper endeavors to evaluate the competency of popular large vision-language models (LVLMs) in specialized and general tasks. We employ six challenging tasks in three different application scenarios: natural, healthcare, and industrial. We examine the performance of three recent open-source LVLMs, including MiniGPT-v2, LLaVA-1.5, and Shikra, on both visual recognition and localization in these tasks.
  arXiv  Detail & Related papers  (2024-03-07T08:25:27Z)
• Competition-Level Problems are Effective LLM Evaluators [121.15880285283116]
  This paper aims to evaluate the reasoning capacities of large language models (LLMs) in solving recent programming problems in Codeforces. We first provide a comprehensive evaluation of GPT-4's peiceived zero-shot performance on this task, considering various aspects such as problems' release time, difficulties, and types of errors encountered. Surprisingly, theThoughtived performance of GPT-4 has experienced a cliff like decline in problems after September 2021 consistently across all the difficulties and types of problems.
  arXiv  Detail & Related papers  (2023-12-04T18:58:57Z)
• Assessing Quality-Diversity Neuro-Evolution Algorithms Performance in Hard Exploration Problems [10.871978893808533]
  Quality-Diversity (QD) methods are evolutionary algorithms inspired by nature's ability to produce high-performing niche organisms. In this paper, we highlight three candidate benchmarks exhibiting control problems in high dimension with exploration difficulties. We also provide open-source implementations in Jax allowing practitioners to run fast and numerous experiments on few compute resources.
  arXiv  Detail & Related papers  (2022-11-24T18:04:12Z)
• Measuring Progress on Scalable Oversight for Large Language Models [19.705153174673576]
  We present an experimental design centered on choosing tasks for which human specialists succeed but unaided humans and current general AI systems fail. We find that human participants who interact with an unreliable large-language-model dialog assistant through chat substantially outperform both the model alone and their own unaided performance.
  arXiv  Detail & Related papers  (2022-11-04T17:03:49Z)
• Learning to Solve Complex Tasks by Talking to Agents [39.08818632689814]
  Humans often solve complex problems by interacting with existing agents, such as AI assistants, that can solve simpler sub-tasks. Common NLP benchmarks aim for the development of self-sufficient models for every task. We propose a new benchmark called CommaQA that contains three kinds of complex reasoning tasks designed to be solved by talking'' to four agents with different capabilities.
  arXiv  Detail & Related papers  (2021-10-16T10:37:34Z)

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.