# Truthful Question Answering on TruthfulQA (Mean per-step regret) Source: [Truthful Question Answering on TruthfulQA (Mean per-step regret)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:52:13.459Z Evaluation of model performance on the TruthfulQA dataset, specifically reporting the Mean per-step regret metric. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Truthful Question Answering | | Dataset (including any stated split) | TruthfulQA | | Metric | Mean Per-Step Regret | | Unit | SCALAR | | Direction | Lower is better | | Results loaded | 15 of 15 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Mean Per-Step Regret](): Lower is better; unit: SCALAR. ## Recorded results | Method | Mean Per-Step Regret | Result date | Report paper | | --- | --- | --- | --- | | ϵ-FTRL | 0.138 | Not recorded | 699e820b6017c4460ffacf00 | | Simpl | 0.141 | Not recorded | 699e820b6017c4460ffacf00 | | LinUCB+KL | 0.15 | Not recorded | 699e820b6017c4460ffacf00 | | No-Rewrite (NoRw) | 0.151 | Not recorded | 699e820b6017c4460ffacf00 | | TS | 0.155 | Not recorded | 699e820b6017c4460ffacf00 | | TS | 0.155 | Not recorded | 699e820b6017c4460ffacf00 | | Para | 0.159 | Not recorded | 699e820b6017c4460ffacf00 | | LinUCB | 0.161 | Not recorded | 699e820b6017c4460ffacf00 | | Disamb | 0.166 | Not recorded | 699e820b6017c4460ffacf00 | | FTPL | 0.167 | Not recorded | 699e820b6017c4460ffacf00 | | LinEXP3 | 0.171 | Not recorded | 699e820b6017c4460ffacf00 | | EXP3 | 0.173 | Not recorded | 699e820b6017c4460ffacf00 | | Clarify | 0.18 | Not recorded | 699e820b6017c4460ffacf00 | | LinFTPL | 0.18 | Not recorded | 699e820b6017c4460ffacf00 | | Expand | 0.206 | Not recorded | 699e820b6017c4460ffacf00 | ### ϵ-FTRL Mean Per-Step Regret: 0.138. Result ID: 699eaa375698665ec4767062. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Non-Contextual"}. This record does not assert independent reproduction. ### Simpl Mean Per-Step Regret: 0.141. Result ID: 699eaa375698665ec476705c. [Reporting paper]() Recorded evaluation configuration: {"Strategy Category":"Static Prompts"}. This record does not assert independent reproduction. ### LinUCB+KL Mean Per-Step Regret: 0.15. Result ID: 699eaa375698665ec4767065. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Contextual Linear"}. This record does not assert independent reproduction. ### No-Rewrite (NoRw) Mean Per-Step Regret: 0.151. Result ID: 699eaa375698665ec476705a. [Reporting paper]() Recorded evaluation configuration: {"Strategy Category":"Base"}. This record does not assert independent reproduction. ### TS Mean Per-Step Regret: 0.155. Result ID: 699eaa375698665ec4767063. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Non-Contextual"}. This record does not assert independent reproduction. ### TS Mean Per-Step Regret: 0.155. Result ID: 699eaa375698665ec4767068. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Contextual Linear"}. This record does not assert independent reproduction. ### Para Mean Per-Step Regret: 0.159. Result ID: 699eaa375698665ec476705b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Static Prompts"}. This record does not assert independent reproduction. ### LinUCB Mean Per-Step Regret: 0.161. Result ID: 699eaa375698665ec4767064. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Contextual Linear"}. This record does not assert independent reproduction. ### Disamb Mean Per-Step Regret: 0.166. Result ID: 699eaa375698665ec476705d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Static Prompts"}. This record does not assert independent reproduction. ### FTPL Mean Per-Step Regret: 0.167. Result ID: 699eaa375698665ec4767061. [Reporting paper]() Recorded evaluation configuration: {"Strategy Category":"Non-Contextual"}. This record does not assert independent reproduction. ### LinEXP3 Mean Per-Step Regret: 0.171. Result ID: 699eaa375698665ec4767066. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Contextual Linear"}. This record does not assert independent reproduction. ### EXP3 Mean Per-Step Regret: 0.173. Result ID: 699eaa375698665ec4767060. [Reporting paper]() Recorded evaluation configuration: {"Strategy Category":"Non-Contextual"}. This record does not assert independent reproduction. ### Clarify Mean Per-Step Regret: 0.18. Result ID: 699eaa375698665ec476705e. [Reporting paper]() Recorded evaluation configuration: {"Strategy Category":"Static Prompts"}. This record does not assert independent reproduction. ### LinFTPL Mean Per-Step Regret: 0.18. Result ID: 699eaa375698665ec4767067. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Contextual Linear"}. This record does not assert independent reproduction. ### Expand Mean Per-Step Regret: 0.206. Result ID: 699eaa375698665ec476705f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Strategy Category":"Static Prompts"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.