CoverBench: A Challenging Benchmark for Complex Claim Verification

TL;DR

CoverBench is a new, challenging benchmark designed to evaluate the ability of language models to verify complex claims across diverse domains and reasoning types, highlighting the need for improved model performance.

Contribution

It introduces a diversified, high-quality dataset for complex claim verification, with manual vetting and multiple representations, filling a gap in existing benchmarks.

Findings

CoverBench is challenging with significant headroom for improvement.

Baseline models perform poorly on complex reasoning tasks.

The dataset covers various domains and reasoning types.

Abstract

There is a growing line of research on verifying the correctness of language models' outputs. At the same time, LMs are being used to tackle complex queries that require reasoning. We introduce CoverBench, a challenging benchmark focused on verifying LM outputs in complex reasoning settings. Datasets that can be used for this purpose are often designed for other complex reasoning tasks (e.g., QA) targeting specific use-cases (e.g., financial tables), requiring transformations, negative sampling and selection of hard examples to collect such a benchmark. CoverBench provides a diversified evaluation for complex claim verification in a variety of domains, types of reasoning, relatively long inputs, and a variety of standardizations, such as multiple representations for tables where available, and a consistent schema. We manually vet the data for quality to ensure low levels of label noise. Finally, we report a variety of competitive baseline results to show CoverBench is challenging and has very significant headroom. The data is available at https://huggingface.co/datasets/google/coverbench .