Hallucination Rate Assessment on HEAL Probes TaskDescMod - DistInj
2.58COLlama-3-8B-Instruct
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Llama-3-8B-InstructEnv=BEHAVIOR2026.04 | 2.58 | 9.49 | |
| Qwen-14B-Instruct-VCDEnv=BEHAVIOR2026.04 | 3.6 | 5 | |
| Qwen-14B-InstructEnv=BEHAVIOR2026.04 | 4.1 | 16.5 | |
| Gemma-2-9b-itEnv=BEHAVIOR2026.04 | 11.6 | 5.7 | |
| DS-R1-Distil-LLaMA-8BEnv=BEHAVIOR2026.04 | 13.9 | 27.2 |