List your company

Hallucination Detection on SQuAD (AUCs, AUCr, PCC Metrics)

69.9AUCsLSC
26.7437.94549.1560.355Jan 7, 2026

Evaluation results

MethodLinks
LSCModel=LLaMA-7B2026.0169.969.40.327—
LSCModel=LLaMA-13B2026.0169.169.70.348—
EigenScoreModel=LLaMA-7B2026.0167.567.80.342—
EigenScoreModel=LLaMA-13B2026.0166.166.40.327—
Lexical SimilarityModel=LLaMA-7B2026.0164.265.40.252—
Lexical SimilarityModel=LLaMA-13B2026.016264.30.234—
LSCModel=Qwen-3B2026.0161.362.10.197—
EigenScoreModel=Qwen-7B2026.0158.5600.142—
EigenScoreModel=Qwen-3B2026.0157.959.50.153—
Lexical SimilarityModel=Qwen-7B2026.0157.259.80.129—
Lexical SimilarityModel=Qwen-3B2026.0155.158.60.12—
AGSERModel=Qwen-3B2026.0154.257.30.122—
LSCModel=Qwen-7B2026.0153.8560.073—
LN-EntropyModel=LLaMA-13B2026.0153.556.40.098—
LN-EntropyModel=LLaMA-7B2026.0153.556.50.072—
AGSERModel=LLaMA-13B2026.0150.851.9-0.015—
LN-EntropyModel=Qwen-3B2026.0149.553.60.048—
LN-EntropyModel=Qwen-7B2026.0148.951.6-0.012—
AGSERModel=Qwen-7B2026.0145.948.8-0.076—
PerplexityModel=Qwen-3B2026.0145.851.40.039—
AGSERModel=LLaMA-7B2026.0143.245.5-0.053—
PerplexityModel=LLaMA-7B2026.0140.646.2-0.055—
PerplexityModel=LLaMA-13B2026.0139.744.8-0.066—
EnergyModel=Qwen-3B2026.0138.942.4-0.07—
PerplexityModel=Qwen-7B2026.0138.342-0.14—
EnergyModel=LLaMA-13B2026.0135.438.5-0.211—
EnergyModel=LLaMA-7B2026.0132.237.4-0.189—
EnergyModel=Qwen-7B2026.0128.430.8-0.345—
HaMI (Original)Backbone=LLaMA-3.1-8B2026.05———79.5
HaMI (Original)Backbone=Mistral-Nemo-Instruct (12B)2026.05———84
HaMI (Original)Backbone=LLaMA-3.3-Instruct-70B2026.05———84.1
HaMI (SP)Backbone=LLaMA-3.1-8B2026.05———84.4
HaMI (SP)Backbone=Mistral-Nemo-Instruct (12B)2026.05———85.8
HaMI (SP)Backbone=LLaMA-3.3-Instruct-70B2026.05———87.4
Max PoolingBackbone=LLaMA-3.1-8B2026.05———82.9
Max PoolingBackbone=Mistral-Nemo-Instruct (12B)2026.05———85.9
Max PoolingBackbone=LLaMA-3.3-Instruct-70B2026.05———87.4
Mean PoolingBackbone=LLaMA-3.1-8B2026.05———81.7
Mean PoolingBackbone=Mistral-Nemo-Instruct (12B)2026.05———84.6
Mean PoolingBackbone=LLaMA-3.3-Instruct-70B2026.05———85.4