Commonsense Reasoning on HellaSwag (ECE, BS)
3.23ECEVerbalized
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VerbalizedLLM=QWEN-2 7B2026.06 | 3.23 | 15.73 | |
| CaliDist_verbalized (As.)LLM=QWEN-2 7B2026.06 | 3.52 | 15.98 | |
| CaliDist_verbalized (Pr.)LLM=QWEN-2 7B2026.06 | 3.52 | 15.98 | |
| CaliDist_verbalized (Sa.)LLM=QWEN-2 7B2026.06 | 3.52 | 15.98 | |
| CaliDist_verbalized (Pr.)LLM=LLAMA-3.1 8B2026.06 | 9.11 | 23.54 | |
| CaliDist_verbalized (Sa.)LLM=LLAMA-3.1 8B2026.06 | 9.12 | 23.54 | |
| CaliDist_verbalized (As.)LLM=LLAMA-3.1 8B2026.06 | 9.15 | 23.55 | |
| VerbalizedLLM=LLAMA-3.1 8B2026.06 | 11.56 | 21.06 |