PACT Benchmark

What is PACT?

PACT evaluates LLM assistants' compliance with standing rules under nine realistic pressures across 48 scenarios and 12 regulated domains, scoring multi-turn…

Released

2026-09-16

Evaluates

General AI, Language & Knowledge, cs.CL

Openness

unknown

Importer

Claire Radar

Review status

ai-reviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2609.18605

PACT paper, code and dataset

• Paper https://arxiv.org/abs/2609.18605
• Code https://github.com/trace-ai-labs/pact
• Dataset https://huggingface.co/datasets/trace-ai-labs/pact

Which sources cite PACT?

• PACT

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

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 2d-fet-bench
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A Real-World CAD Segmentation Benchmark
• A3S-Bench
• AARRI-Bench