The Automated LLM Speedrunning Benchmark: Reproducing NanoGPT Improvements

• URL: http://arxiv.org/abs/2506.22419v2
• Date: Mon, 30 Jun 2025 21:56:29 GMT
• Title: The Automated LLM Speedrunning Benchmark: Reproducing NanoGPT Improvements
• Authors: Bingchen Zhao, Despoina Magka, Minqi Jiang, Xian Li, Roberta Raileanu, Tatiana Shavrina, Jean-Christophe Gagnon-Audet, Kelvin Niu, Shagun Sodhani, Michael Shvartsman, Andrei Lupu, Alisia Lupidi, Edan Toledo, Karen Hambardzumyan, Martin Josifoski, Thomas Foster, Lucia Cipolina-Kun, Abhishek Charnalia, Derek Dunfield, Alexander H. Miller, Oisin Mac Aodha, Jakob Foerster, Yoram Bachrach,
• Abstract summary: A critical capability toward scientific progress is the ability to reproduce existing work.<n>To evaluate the ability of AI agents to reproduce results in an active research area, we introduce the Automated LLM Speedrunning Benchmark.<n>We find that recent reasoning LLMs combined with SoTA scaffolds struggle to reimplement already-known innovations in our benchmark.
• Score: 87.61432174951891
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Rapid advancements in large language models (LLMs) have the potential to assist in scientific progress. A critical capability toward this endeavor is the ability to reproduce existing work. To evaluate the ability of AI agents to reproduce results in an active research area, we introduce the Automated LLM Speedrunning Benchmark, leveraging the research community contributions on the NanoGPT speedrun, a competition to train a GPT-2 model in the shortest time. Each of the 19 speedrun tasks provides the agent with the previous records training script, optionally paired with one of three hint formats, ranging from pseudocode to paper-like descriptions of the new records improvements. Records execute quickly by design and speedrun improvements encompass diverse code-level changes, ranging from high-level algorithmic advancements to hardware-aware optimizations. These features make the benchmark both accessible and realistic for the frontier problem of improving LLM training. We find that recent reasoning LLMs combined with SoTA scaffolds struggle to reimplement already-known innovations in our benchmark, even when given detailed hints. Our benchmark thus provides a simple, non-saturated measure of an LLMs ability to automate scientific reproduction, a necessary (but not sufficient) skill for an autonomous research agent.

Related papers

• Assistax: A Hardware-Accelerated Reinforcement Learning Benchmark for Assistive Robotics [18.70896736010314]
  Games have dominated reinforcement learning benchmarks because they present relevant challenges, are inexpensive to run and easy to understand.<n>We introduce Assistax: an open-source benchmark designed to address challenges arising in assistive robotics tasks.<n>In terms of open-loop wall-clock time, Assistax runs up to $370times$ faster when vectorising training runs compared to CPU-based alternatives.
  arXiv  Detail & Related papers  (2025-07-29T09:49:11Z)
• Grammar-Guided Evolutionary Search for Discrete Prompt Optimisation [63.97051732013936]
  We propose an evolutionary search approach to automated discrete prompt optimisation consisting of two phases.<n>In the first phase, grammar-guided genetic programming is invoked to synthesise prompt-creating programmes.<n>In the second phase, local search is applied to explore the neighbourhoods of best-performing programmes.
  arXiv  Detail & Related papers  (2025-07-14T14:34:15Z)
• MDBench: A Synthetic Multi-Document Reasoning Benchmark Generated with Knowledge Guidance [5.192956837901584]
  We introduce MDBench, a new dataset for evaluating large language mod-els (LLMs) on the task of multi-document reasoning.<n>We use a novel synthetic generation process, allowing us to controllably and efficiently generate challenging document sets.<n>We analyze the behavior of popular LLMs and prompting techniques, finding that MDBENCH poses significant challenges for all methods.
  arXiv  Detail & Related papers  (2025-06-17T19:14:30Z)
• Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognition [95.54406667705999]
  Pangu Embedded is an efficient Large Language Model (LLM) reasoner developed on Ascend Neural Processing Units (NPUs)<n>It addresses the significant computational costs and inference latency challenges prevalent in existing reasoning-optimized LLMs.<n>It delivers rapid responses and state-of-the-art reasoning quality within a single, unified model architecture.
  arXiv  Detail & Related papers  (2025-05-28T14:03:02Z)
• Iterative Self-Incentivization Empowers Large Language Models as Agentic Searchers [74.17516978246152]
  Large language models (LLMs) have been widely integrated into information retrieval to advance traditional techniques.<n>We propose EXSEARCH, an agentic search framework, where the LLM learns to retrieve useful information as the reasoning unfolds.<n>Experiments on four knowledge-intensive benchmarks show that EXSEARCH substantially outperforms baselines.
  arXiv  Detail & Related papers  (2025-05-26T15:27:55Z)
• CodeIF: Benchmarking the Instruction-Following Capabilities of Large Language Models for Code Generation [24.090719826360342]
  We introduce CodeIF, the first benchmark designed to assess the abilities of Large Language Models (LLMs) to adhere to task-oriented instructions within code generation scenarios.<n>We conduct extensive experiments with LLMs, analyzing their strengths and limitations in meeting the demands of these tasks.
  arXiv  Detail & Related papers  (2025-02-26T14:19:49Z)
• 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)
• LLM Program Optimization via Retrieval Augmented Search [71.40092732256252]
  We propose a blackbox adaptation method called Retrieval Augmented Search (RAS) that performs beam search over candidate optimizations.<n>We show that RAS performs 1.8$times$ better than prior state-of-the-art blackbox adaptation strategies.<n>We also propose a method called AEGIS for improving interpretability by decomposing training examples into "atomic edits"
  arXiv  Detail & Related papers  (2025-01-31T06:34:47Z)
• Attention in Large Language Models Yields Efficient Zero-Shot Re-Rankers [7.6245627565464]
  Large language models (LLMs) have become popular choices for zero-shot re-ranking in IR systems.<n>We propose in-context re-ranking (ICR), a novel method that leverages the change in attention pattern caused by the search query for accurate and efficient re-ranking.<n>Our findings call for further exploration on novel ways of utilizing open-weight LLMs beyond text generation.
  arXiv  Detail & Related papers  (2024-10-03T16:25:37Z)
• Accelerating LLaMA Inference by Enabling Intermediate Layer Decoding via Instruction Tuning with LITE [62.13435256279566]
  Large Language Models (LLMs) have achieved remarkable performance across a wide variety of natural language tasks. However, their large size makes their inference slow and computationally expensive. We show that it enables these layers to acquire 'good' generation ability without affecting the generation ability of the final layer.
  arXiv  Detail & Related papers  (2023-10-28T04:07:58Z)
• 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)
• Prompts Matter: Insights and Strategies for Prompt Engineering in Automated Software Traceability [45.235173351109374]
  Large Language Models (LLMs) have the potential to revolutionize automated traceability. This paper explores the process of prompt engineering to extract link predictions from an LLM.
  arXiv  Detail & Related papers  (2023-08-01T01:56:22Z)

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.