Consistent Reasoning on TruthfulQA
79.4AccuracyGibbs
Evaluation results
| Method | Links | |
|---|---|---|
| GibbsLLM=Llama-3.1-8B-Instruct, n=642026.06 | 79.4 | |
| GibbsLLM=OLMo-3-32B-Think, n=162026.06 | 78.8 | |
| GibbsLLM=Llama-3.1-8B-Instruct, n=42026.06 | 77.3 | |
| GibbsLLM=Llama-3.1-8B-Instruct, n=162026.06 | 77.1 | |
| GibbsLLM=OLMo-3-32B-Think, n=642026.06 | 76.6 | |
| 4-Pass (autoreg.)LLM=Llama-3.1-8B-Instruct2026.06 | 75.9 | |
| ICMLLM=Llama-3.1-8B-Instruct, n=642026.06 | 75.9 | |
| GibbsLLM=OLMo-3-32B-Think, n=42026.06 | 75.9 | |
| 1-Pass (autoreg.)LLM=Llama-3.1-8B-Instruct2026.06 | 73.4 | |
| ICMLLM=OLMo-3-32B-Think, n=642026.06 | 73.2 | |
| Zero-shot (greedy)LLM=OLMo-3-32B-Think2026.06 | 68.3 | |
| 4-Pass (autoreg.)LLM=OLMo-3-32B-Think2026.06 | 67.6 | |
| Zero-shot (greedy)LLM=Llama-3.1-8B-Instruct2026.06 | 65.7 | |
| 1-Pass (autoreg.)LLM=OLMo-3-32B-Think2026.06 | 64.9 |