Open-ended Generation on TruthfulQA single info 1.0 (test)
66.6Truthfulness ScoreTACS-T
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TACS-TBackbone=Mistral-Instruct-v0.2, Selection Strategy=Token-level Truth-Aware Context Selection2024.03 | 66.6 | 58 | |
| TACS-SBackbone=Mistral-Instruct-v0.2, Selection Strategy=Sentence-level Truth-Aware Context Selection2024.03 | 61.8 | 55.2 | |
| Mistral-Instruct-v0.2Backbone=Mistral-Instruct-v0.2, Selection Strategy=Baseline2024.03 | 59.9 | 52.7 | |
| TACS-SBackbone=Llama 2-Chat, Selection Strategy=Sentence-level Truth-Aware Context Selection2024.03 | 59.4 | 55.4 | |
| TACS-TBackbone=Llama 2-Chat, Selection Strategy=Token-level Truth-Aware Context Selection2024.03 | 56.9 | 53.2 | |
| Llama 2-ChatBackbone=Llama 2-Chat, Selection Strategy=Baseline2024.03 | 55.1 | 51.6 | |
| ITIBackbone=Llama 2-Chat, Selection Strategy=Inference-Time Intervention2024.03 | 53.2 | 49.9 |