Can LLM Agents Discover? Evaluating Creativity on ML Engineering Tasks

• URL: http://arxiv.org/abs/2608.30047v1
• Date: Sun, 30 Aug 2026 21:21:08 GMT
• Title: Can LLM Agents Discover? Evaluating Creativity on ML Engineering Tasks
• Abstract summary: We present a framework for evaluating multi-turn LLM research agents' creativity using ML engineering tasks as a testbed.<n>We develop an LLM-as-a-Judge pipeline and verify its strong correlation with human creativity judgments.<n>Applying this pipeline to agent trajectories, we find: (1) all agents exhibit declining P-Creativity as they transition from exploration to exploitation; (2) LLMs exhibit greater H-Creativity than medal-winning humans, yet achieve lower performance.
• Score: 6.696727578876306
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent AI systems promise autonomous scientific discovery, claiming to discover algorithms and produce research papers, yet understanding whether they exhibit creativity, the capacity to produce solutions that are both novel and useful, remains an open question. We present a framework for evaluating multi-turn LLM research agents' creativity using ML engineering tasks as a testbed, through three dimensions: P-Creativity (psychological novelty: novel relative to the agent's own prior solutions within a run), H-Creativity (historical novelty: novel relative to the corpus of human solutions), and Usefulness (task performance). Evaluating two agent frameworks, AIDE and AIRA-Dojo, on 10 Kaggle-style machine learning tasks from MLE-Bench, we develop an LLM-as-a-Judge pipeline and verify its strong correlation with human creativity judgments, providing a reliable automated metric for P-Creativity evaluation at scale. Applying this pipeline to agent trajectories, we find: (1) all agents exhibit declining P-Creativity as they transition from exploration to exploitation; (2) LLMs exhibit greater H-Creativity than medal-winning humans, yet achieve lower performance. Our findings reveal that current agents can explore novel regions of the solution space but lack the capacity to convert this novelty into improved task performance.

Related papers

• Automated Creativity Evaluation of Language Models Across Open-Ended Tasks [3.6948012022450776]
  Large language models (LLMs) have achieved remarkable progress in language understanding, reasoning, and generation.<n>We introduce an automated, domain-agnostic framework for quantifying LLM creativity across open-ended tasks.<n>Our approach separates the measurement apparatus from the creative task itself, enabling scalable, task-agnostic assessment.
  arXiv  Detail & Related papers  (2026-06-10T07:37:06Z)
• LLM Agents Beyond Utility: An Open-Ended Perspective [50.809163251551894]
  We augment a pretrained LLM agent with the ability to generate its own tasks, accumulate knowledge, and interact extensively with its environment.<n>It can reliably follow complex multi-step instructions, store and reuse information across runs, and propose and solve its own tasks.<n>It remains sensitive to prompt design, prone to repetitive task generation, and unable to form self-representations.
  arXiv  Detail & Related papers  (2025-10-16T10:46:54Z)
• 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)
• Agentic Additive Manufacturing Alloy Discovery [10.771896497348969]
  This work aims to utilize Large Language Model (LLM) enabled agents to automate and accelerate the task of alloy discovery within the field of additive manufacturing.
  arXiv  Detail & Related papers  (2025-10-02T21:06:04Z)
• ReMA: Learning to Meta-think for LLMs with Multi-Agent Reinforcement Learning [53.817538122688944]
  We introduce Reinforced Meta-thinking Agents (ReMA) to elicit meta-thinking behaviors from Reasoning of Large Language Models (LLMs)<n>ReMA decouples the reasoning process into two hierarchical agents: a high-level meta-thinking agent responsible for generating strategic oversight and plans, and a low-level reasoning agent for detailed executions.<n> Empirical results from single-turn experiments demonstrate that ReMA outperforms single-agent RL baselines on complex reasoning tasks.
  arXiv  Detail & Related papers  (2025-03-12T16:05:31Z)
• MLGym: A New Framework and Benchmark for Advancing AI Research Agents [51.9387884953294]
  We introduce Meta MLGym and MLGym-Bench, a new framework and benchmark for evaluating and developing large language models on AI research tasks.<n>This is the first Gym environment for machine learning (ML) tasks, enabling research on reinforcement learning (RL) algorithms for training such agents.<n>We evaluate a number of frontier large language models (LLMs) on our benchmarks such as Claude-3.5-Sonnet, Llama-3.1 405B, GPT-4o, o1-preview, and Gemini-1.5 Pro.
  arXiv  Detail & Related papers  (2025-02-20T12:28:23Z)
• Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers [90.26363107905344]
  Large language models (LLMs) have sparked optimism about their potential to accelerate scientific discovery. No evaluations have shown that LLM systems can take the very first step of producing novel, expert-level ideas.
  arXiv  Detail & Related papers  (2024-09-06T08:25:03Z)
• Benchmarking Language Model Creativity: A Case Study on Code Generation [39.546827184857754]
  In this work, we introduce a framework for quantifying LLM creativity.<n>We define NEOGAUGE, a metric that quantifies both convergent and divergent thinking in the generated creative responses.<n>We test the proposed framework on Codeforces problems, which serve as both a natural dataset for coding tasks and a collection of prior human solutions.
  arXiv  Detail & Related papers  (2024-07-12T05:55:22Z)
• WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks [85.95607119635102]
  Large language models (LLMs) can mimic human-like intelligence.<n>WorkArena++ is designed to evaluate the planning, problem-solving, logical/arithmetic reasoning, retrieval, and contextual understanding abilities of web agents.
  arXiv  Detail & Related papers  (2024-07-07T07:15:49Z)
• Luban: Building Open-Ended Creative Agents via Autonomous Embodied Verification [34.97881486372797]
  Building open agents has always been the ultimate goal in AI research, and creative agents are the more enticing. We introduce autonomous embodied verification techniques for agents to fill the gap, laying the groundwork for creative tasks. Specifically, we propose the Luban agent target creative building tasks in Minecraft, which equips with two-level autonomous embodied verification.
  arXiv  Detail & Related papers  (2024-05-24T10:25:59Z)
• MacGyver: Are Large Language Models Creative Problem Solvers? [87.70522322728581]
  We explore the creative problem-solving capabilities of modern LLMs in a novel constrained setting.<n>We create MACGYVER, an automatically generated dataset consisting of over 1,600 real-world problems.<n>We present our collection to both LLMs and humans to compare and contrast their problem-solving abilities.
  arXiv  Detail & Related papers  (2023-11-16T08:52:27Z)
• MLAgentBench: Evaluating Language Agents on Machine Learning Experimentation [96.71370747681078]
  We introduce MLAgentBench, a suite of 13 tasks ranging from improving model performance on CIFAR-10 to recent research problems like BabyLM. For each task, an agent can perform actions like reading/writing files, executing code, and inspecting outputs. We benchmark agents based on Claude v1.0, Claude v2.1, Claude v3 Opus, GPT-4, GPT-4-turbo, Gemini-Pro, and Mixtral and find that a Claude v3 Opus agent is the best in terms of success rate.
  arXiv  Detail & Related papers  (2023-10-05T04:06:12Z)

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.