DeFAb Benchmark

What is DeFAb?

A formally grounded benchmark and generation pipeline for evaluating defeasible abduction through automatically verifiable hypothesis construction.

Released

2026-06-17

Evaluates

Factuality, General AI, Language & Knowledge, cs.AI

Openness

unknown

Importer

Claire Radar

Review status

ai-reviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2606.18557

DeFAb paper, code and dataset

• Paper https://arxiv.org/abs/2606.18557
• Code https://github.com/PatrickAllenCooper/blanc
• Dataset https://huggingface.co/datasets/PatrickAllenCooper/DeFAb

Which sources cite DeFAb?

• DeFAb

Related Factuality, General AI, Language & Knowledge, cs.AI benchmarks

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