The BiGGen Bench: A Principled Benchmark for Fine-grained Evaluation of Language Models with Language Models

TL;DR

The BiGGen Bench provides a detailed, capability-specific evaluation framework for language models, addressing limitations of existing benchmarks by using instance-specific criteria across diverse tasks.

Contribution

It introduces a novel, principled benchmark with instance-specific evaluation for comprehensive, fine-grained assessment of language models' capabilities.

Findings

Evaluated 103 language models using the benchmark.

Demonstrated the effectiveness of instance-specific evaluation criteria.

Provided publicly available code and data for reproducibility.

Abstract

As language models (LMs) become capable of handling a wide range of tasks, their evaluation is becoming as challenging as their development. Most generation benchmarks currently assess LMs using abstract evaluation criteria like helpfulness and harmlessness, which often lack the flexibility and granularity of human assessment. Additionally, these benchmarks tend to focus disproportionately on specific capabilities such as instruction following, leading to coverage bias. To overcome these limitations, we introduce the BiGGen Bench, a principled generation benchmark designed to thoroughly evaluate nine distinct capabilities of LMs across 77 diverse tasks. A key feature of the BiGGen Bench is its use of instance-specific evaluation criteria, closely mirroring the nuanced discernment of human evaluation. We apply this benchmark to assess 103 frontier LMs using five evaluator LMs. Our code, data, and evaluation results are all publicly available at https://github.com/prometheus-eval/prometheus-eval/tree/main/BiGGen-Bench.