The Automated LLM Speedrunning Benchmark: Reproducing NanoGPT Improvements

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 Maria Lupidi, Karen Hambardzumyan, Martin Josifoski, Edan Toledo, Thomas Foster, Lucia Cipolina-Kun, Derek Dunfield, Abhishek Charnalia, Alexander H Miller et al. (3 additional authors not shown)

Keywords: LLM agents, automated scientific discovery, scientific reproducibility, code generation, AI agent scaffolding, LLM pre-training

TL;DR: We introduce the Automated LLM Speedrunning benchmark to assess the capabilities of AI agents to reproduce LLM research

Abstract: Rapidly improving large language models (LLMs) have the potential to assist in scientific progress. One critical skill in this endeavor is the ability to faithfully reproduce existing work. To evaluate the capability of AI agents to reproduce complex code in an active research area, we introduce the Automated LLM Speedrunning Benchmark, leveraging the research community's contributions to the $\textit{NanoGPT speedrun}$, a competition to train a GPT-2 model in the shortest time. Each of the 19 speedrun tasks provides the agent with the previous record's training script, optionally paired with one of three hint formats, ranging from pseudocode to paper-like descriptions of the new record's 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 frontier 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 LLM's ability to automate scientific reproduction, a necessary (but not sufficient) skill for an autonomous research agent.

Dataset URL: https://github.com/facebookresearch/llm-speedrunner

Code URL: https://github.com/facebookresearch/llm-speedrunner

Primary Area: Datasets & Benchmarks for applications in language modeling and vision language modeling

Submission Number: 822