GPT-Fathom: Benchmarking Large Language Models to Decipher the Evolutionary Path towards GPT-4 and Beyond

TL;DR

GPT-Fathom provides a comprehensive, reproducible benchmark suite for evaluating large language models, offering insights into their evolution from GPT-3 to GPT-4 and highlighting factors influencing their capabilities.

Contribution

It introduces GPT-Fathom, an open-source evaluation framework that systematically compares multiple LLMs under consistent settings, addressing limitations of previous leaderboards.

Findings

OpenAI's models show progressive improvements from GPT-3 to GPT-4.

Adding code data enhances reasoning capabilities of LLMs.

Alignment techniques like SFT and RLHF impact model performance and alignment tax.

Abstract

With the rapid advancement of large language models (LLMs), there is a pressing need for a comprehensive evaluation suite to assess their capabilities and limitations. Existing LLM leaderboards often reference scores reported in other papers without consistent settings and prompts, which may inadvertently encourage cherry-picking favored settings and prompts for better results. In this work, we introduce GPT-Fathom, an open-source and reproducible LLM evaluation suite built on top of OpenAI Evals. We systematically evaluate 10+ leading LLMs as well as OpenAI's legacy models on 20+ curated benchmarks across 7 capability categories, all under aligned settings. Our retrospective study on OpenAI's earlier models offers valuable insights into the evolutionary path from GPT-3 to GPT-4. Currently, the community is eager to know how GPT-3 progressively improves to GPT-4, including technical details like whether adding code data improves LLM's reasoning capability, which aspects of LLM capability can be improved by SFT and RLHF, how much is the alignment tax, etc. Our analysis sheds light on many of these questions, aiming to improve the transparency of advanced LLMs.