Question Answering on TriviaQA (T, U, F, Rely)
76.8T ScoreBinary
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| BinaryBase Model=Llama3.1-8B-Inst2026.04 | 76.8 | 0 | 23.2 | 76.8 | |
| PromptingBase Model=Llama3.1-8B-Inst2026.04 | 71.6 | 1.3 | 27.1 | 72.9 | |
| SFTBase Model=Llama3.1-8B-Inst2026.04 | 70.2 | 0 | 29.8 | 70.2 | |
| KARLBase Model=Llama3.1-8B-Inst2026.04 | 67.9 | 23.5 | 8.6 | 85.9 | |
| BinaryBase Model=Qwen2.5-7B-Inst2026.04 | 64.2 | 0 | 35.8 | 64.2 | |
| Ternary (TruthRL)Base Model=Llama3.1-8B-Inst2026.04 | 61.4 | 32.3 | 6.3 | 83.3 | |
| SFTBase Model=Qwen2.5-7B-Inst2026.04 | 60.3 | 0 | 39.7 | 60.3 | |
| RLKFBase Model=Llama3.1-8B-Inst2026.04 | 57.6 | 23.2 | 19.2 | 75.4 | |
| R-TuningBase Model=Llama3.1-8B-Inst2026.04 | 55.4 | 35.9 | 8.7 | 78.4 | |
| KARLBase Model=Qwen2.5-7B-Inst2026.04 | 51.7 | 30.9 | 17.4 | 73.1 | |
| PromptingBase Model=Qwen2.5-7B-Inst2026.04 | 49.7 | 17 | 33.3 | 63.8 | |
| RLKFBase Model=Qwen2.5-7B-Inst2026.04 | 41.7 | 1.7 | 56.6 | 43.4 | |
| Ternary (TruthRL)Base Model=Qwen2.5-7B-Inst2026.04 | 39.2 | 52.2 | 8.6 | 64.2 | |
| R-TuningBase Model=Qwen2.5-7B-Inst2026.04 | 35.4 | 60.1 | 4.5 | 59.4 |