Factuality Question Answering on TruthfulQA (score)
66.92ScoreUNA-score & binary
Evaluation Results
| Method | Links | |
|---|---|---|
| UNA-score & binaryBase Model=Qwen 8B, Alignment Strategy=UNA-score & binary, Feedback Type=Score-based & Binary2024.08 | 66.92 | |
| UNA-score & binaryBase Model=Mistral 7B, Alignment Strategy=UNA-score & binary, Feedback Type=Score-based & Binary2024.08 | 64.62 | |
| UNA-score (MSE)Base Model=Qwen 8B, Alignment Strategy=UNA-score (MSE), Feedback Type=Score-based2024.08 | 64.44 | |
| UNA-score & binaryBase Model=Llama 8B, Alignment Strategy=UNA-score & binary, Feedback Type=Score-based & Binary2024.08 | 61.24 | |
| UNA-score (MSE)Base Model=Llama 8B, Alignment Strategy=UNA-score (MSE), Feedback Type=Score-based2024.08 | 60.46 | |
| UNA-score (MSE)Base Model=Mistral 7B, Alignment Strategy=UNA-score (MSE), Feedback Type=Score-based2024.08 | 55.09 | |
| KTOBase Model=Llama 8B, Alignment Strategy=KTO, Feedback Type=Binary2024.08 | 55.07 | |
| UNA-binary (BCE)Base Model=Llama 8B, Alignment Strategy=UNA-binary (BCE), Feedback Type=Binary2024.08 | 54.75 | |
| DPOBase Model=Llama 8B, Alignment Strategy=DPO, Feedback Type=Pairwise2024.08 | 53.47 | |
| UNA-pairwiseBase Model=Llama 8B, Alignment Strategy=UNA-pairwise, Feedback Type=Pairwise2024.08 | 53.47 | |
| Qwen 8BBase Model=Qwen 8B, Alignment Strategy=Baseline2024.08 | 52.19 | |
| UNA-binary (BCE)Base Model=Qwen 8B, Alignment Strategy=UNA-binary (BCE), Feedback Type=Binary2024.08 | 52.1 | |
| KTOBase Model=Qwen 8B, Alignment Strategy=KTO, Feedback Type=Binary2024.08 | 52.07 | |
| DPOBase Model=Qwen 8B, Alignment Strategy=DPO, Feedback Type=Pairwise2024.08 | 51.78 | |
| UNA-pairwiseBase Model=Qwen 8B, Alignment Strategy=UNA-pairwise, Feedback Type=Pairwise2024.08 | 51.78 | |
| UNA-binary (BCE)Base Model=Mistral 7B, Alignment Strategy=UNA-binary (BCE), Feedback Type=Binary2024.08 | 48.33 | |
| KTOBase Model=Mistral 7B, Alignment Strategy=KTO, Feedback Type=Binary2024.08 | 47.72 | |
| UNA-score & binaryBase Model=Gemma 4B, Alignment Strategy=UNA-score & binary, Feedback Type=Score-based & Binary2024.08 | 46.69 | |
| UNA-score (MSE)Base Model=Gemma 4B, Alignment Strategy=UNA-score (MSE), Feedback Type=Score-based2024.08 | 45.74 | |
| Llama 8BBase Model=Llama 8B, Alignment Strategy=Baseline2024.08 | 45.16 | |
| DPOBase Model=Mistral 7B, Alignment Strategy=DPO, Feedback Type=Pairwise2024.08 | 44.75 | |
| UNA-pairwiseBase Model=Mistral 7B, Alignment Strategy=UNA-pairwise, Feedback Type=Pairwise2024.08 | 44.75 | |
| Mistral 7BBase Model=Mistral 7B, Alignment Strategy=Baseline2024.08 | 42.58 | |
| UNA-binary (BCE)Base Model=Gemma 4B, Alignment Strategy=UNA-binary (BCE), Feedback Type=Binary2024.08 | 40.79 | |
| DPOBase Model=Gemma 4B, Alignment Strategy=DPO, Feedback Type=Pairwise2024.08 | 40.37 | |
| UNA-pairwiseBase Model=Gemma 4B, Alignment Strategy=UNA-pairwise, Feedback Type=Pairwise2024.08 | 40.37 | |
| Gemma 4BBase Model=Gemma 4B, Alignment Strategy=Baseline2024.08 | 39.73 | |
| KTOBase Model=Gemma 4B, Alignment Strategy=KTO, Feedback Type=Binary2024.08 | 39.71 |