HalluMix is a 6.5k-example, multi-domain benchmark for hallucination detection built from NLI, QA, and summarization data; the authors report Quotient Detections as best with 0.82 accuracy and 0.84 F1.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
HalluMix: A Task-Agnostic, Multi-Domain Benchmark for Real-World Hallucination Detection
HalluMix is a 6.5k-example, multi-domain benchmark for hallucination detection built from NLI, QA, and summarization data; the authors report Quotient Detections as best with 0.82 accuracy and 0.84 F1.