Natural Language Inference on FOLIO
0.61Accuracygemma-2-27b
Evaluation results
| Method | Links | |
|---|---|---|
| gemma-2-27bDecoding Method=GCD2026.01 | 0.61 | |
| gemma-2-9bDecoding Method=GCD2026.01 | 0.53 | |
| gemma-2-27bDecoding Method=NL2LOGIC2026.01 | 0.4 | |
| gemma-2-9bDecoding Method=NL2LOGIC2026.01 | 0.38 | |
| llama-3.2-1bDecoding Method=NL2LOGIC2026.01 | 0.38 | |
| mistral-22bDecoding Method=NL2LOGIC2026.01 | 0.38 | |
| llama-3.1-8bDecoding Method=NL2LOGIC2026.01 | 0.37 | |
| ministral-8bDecoding Method=NL2LOGIC2026.01 | 0.37 | |
| qwen-2.5-7bDecoding Method=NL2LOGIC2026.01 | 0.37 | |
| qwen-2.5-14bDecoding Method=NL2LOGIC2026.01 | 0.37 | |
| llama-3.2-3bDecoding Method=NL2LOGIC2026.01 | 0.36 | |
| gemma-2-2bDecoding Method=NL2LOGIC2026.01 | 0.35 | |
| qwen-2.5-1.5bDecoding Method=NL2LOGIC2026.01 | 0.34 | |
| qwen-2.5-14bDecoding Method=GCD2026.01 | 0.34 | |
| qwen-2.5-3bDecoding Method=NL2LOGIC2026.01 | 0.32 | |
| gemma-2-2bDecoding Method=GCD2026.01 | 0.27 | |
| qwen-2.5-0.5bDecoding Method=NL2LOGIC2026.01 | 0.26 | |
| mistral-22bDecoding Method=GCD2026.01 | 0.23 | |
| qwen-2.5-1.5bDecoding Method=GCD2026.01 | 0.19 | |
| llama-3.2-3bDecoding Method=GCD2026.01 | 0.18 | |
| llama-3.2-1bDecoding Method=GCD2026.01 | 0.17 | |
| qwen-2.5-0.5bDecoding Method=GCD2026.01 | 0.17 | |
| ministral-8bDecoding Method=GCD2026.01 | 0.16 | |
| llama-3.1-8bDecoding Method=GCD2026.01 | 0.12 | |
| qwen-2.5-3bDecoding Method=GCD2026.01 | 0.08 | |
| qwen-2.5-7bDecoding Method=GCD2026.01 | 0.01 |