GRACE Benchmark

What is GRACE?

GRACE evaluates step-level faithfulness of chain-of-thought reasoning in context-grounded tasks, with human annotations and a taxonomy of error categories.

Released

2026-06-15

Evaluates

General AI, Language & Knowledge, Reasoning

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.16151

GRACE paper

• Paper https://arxiv.org/abs/2606.16151

Which sources cite GRACE?

• GRACE

Related General AI, Language & Knowledge, Reasoning benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A-OKVQA
• A3S-Bench
• AARRI-Bench
• ABC-Bench