Hallucination Rate Assessment on HEAL Probes TaskDescMod - TaskObjRem
3.5CO ScoreQwen-14B-Instruct
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen-14B-InstructEnv=BEHAVIOR2026.04 | 3.5 | 10 | |
| Qwen-14B-Instruct-VCDEnv=BEHAVIOR2026.04 | 4.7 | 7.2 | |
| Llama-3-8B-InstructEnv=BEHAVIOR2026.04 | 5 | 6.7 | |
| DS-R1-Distil-LLaMA-8BEnv=BEHAVIOR2026.04 | 6.8 | 13.3 | |
| Gemma-2-9b-itEnv=BEHAVIOR2026.04 | 10.8 | 4.8 |