# Science Question Answering on ARC (Accuracy) Source: [Science Question Answering on ARC (Accuracy)]() Collection index: [llms.txt]() Data updated: 2026-07-07T07:06:00.294Z Evaluation of model performance on the Abstract Reasoning Corpus (ARC) for Science Question Answering, reporting the standard Accuracy score. 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 | Science Question Answering | | Dataset (including any stated split) | ARC | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 13 of 13 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | TTT+CT-KV | 25.8 | Not recorded | 68fa79d51e209131cea7f986 | | TTT | 23.8 | Not recorded | 68fa79d51e209131cea7f986 | | CT-KV | 23.8 | Not recorded | 68fa79d51e209131cea7f986 | | CT-Prompt | 22.5 | Not recorded | 68fa79d51e209131cea7f986 | | Prefix Tuning | 20.5 | Not recorded | 68fa79d51e209131cea7f986 | | Prompt Tuning | 14.5 | Not recorded | 68fa79d51e209131cea7f986 | | LoRA | 13.5 | Not recorded | 68fa79d51e209131cea7f986 | | ICL | 13.3 | Not recorded | 68fa79d51e209131cea7f986 | | DoRA | 13 | Not recorded | 68fa79d51e209131cea7f986 | | Rank-Stabilized LoRA | 12.8 | Not recorded | 68fa79d51e209131cea7f986 | | Prompt Tuning | 12 | Not recorded | 68fa79d51e209131cea7f986 | | Prefix Tuning | 9.3 | Not recorded | 68fa79d51e209131cea7f986 | | Zero-Shot | 1 | Not recorded | 68fa79d51e209131cea7f986 | ### TTT+CT-KV Accuracy: 25.8. Result ID: 6a4ca55760b88bb339c72d25. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"63"}. This record does not assert independent reproduction. ### TTT Accuracy: 23.8. Result ID: 6a4ca55760b88bb339c72d22. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"56"}. This record does not assert independent reproduction. ### CT-KV Accuracy: 23.8. Result ID: 6a4ca55760b88bb339c72d24. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"26"}. This record does not assert independent reproduction. ### CT-Prompt Accuracy: 22.5. Result ID: 6a4ca55760b88bb339c72d23. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"52"}. This record does not assert independent reproduction. ### Prefix Tuning Accuracy: 20.5. Result ID: 6a4ca55760b88bb339c72d21. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"24","m":"# demo"}. This record does not assert independent reproduction. ### Prompt Tuning Accuracy: 14.5. Result ID: 6a4ca55760b88bb339c72d1f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"49","m":"# demo"}. This record does not assert independent reproduction. ### LoRA Accuracy: 13.5. Result ID: 6a4ca55760b88bb339c72d1b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Training time per task (s)":"14"}. This record does not assert independent reproduction. ### ICL Accuracy: 13.3. Result ID: 6a4ca55760b88bb339c72d1a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Training time per task (s)":"0","Evaluation protocol":"In-Context Learning"}. This record does not assert independent reproduction. ### DoRA Accuracy: 13. Result ID: 6a4ca55760b88bb339c72d1d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Training time per task (s)":"15"}. This record does not assert independent reproduction. ### Rank-Stabilized LoRA Accuracy: 12.8. Result ID: 6a4ca55760b88bb339c72d1c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"15"}. This record does not assert independent reproduction. ### Prompt Tuning Accuracy: 12. Result ID: 6a4ca55760b88bb339c72d1e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"13","m":"32"}. This record does not assert independent reproduction. ### Prefix Tuning Accuracy: 9.3. Result ID: 6a4ca55760b88bb339c72d20. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training time per task (s)":"14","m":"32"}. This record does not assert independent reproduction. ### Zero-Shot Accuracy: 1. Result ID: 6a4ca55760b88bb339c72d19. [Reporting paper]() Recorded evaluation configuration: {"Training time per task (s)":"0","Evaluation protocol":"Zero-Shot"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.