# Commonsense Reasoning on ARC-C, ARC-E, PIQA, StoryCloze Source: [Commonsense Reasoning on ARC-C, ARC-E, PIQA, StoryCloze]() Collection index: [llms.txt]() Data updated: 2026-06-02T08:15:54.171Z Evaluation of model performance across a suite of four common sense reasoning benchmarks: ARC-Challenge, ARC-Easy, PIQA, and StoryCloze. 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 | Commonsense Reasoning | | Dataset (including any stated split) | Common Sense Reasoning ARC-C, ARC-E, PIQA, StoryCloze | | Metric | Average Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 34 of 34 | | 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 - [Average Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Average Accuracy | Result date | Report paper | | --- | --- | --- | --- | | ProjQ | 55.58 | Not recorded | 6a1e63bedbd7150506f85338 | | ProjQ | 55.05 | Not recorded | 6a1e63bedbd7150506f85338 | | CALDERA | 54.02 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 53.82 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 53.42 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 52.97 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 49.77 | Not recorded | 6a1e63bedbd7150506f85338 | | ProjQ | 47.82 | Not recorded | 6a1e63bedbd7150506f85338 | | CALDERA | 47.78 | Not recorded | 6a1e63bedbd7150506f85338 | | ProjQ | 47.68 | Not recorded | 6a1e63bedbd7150506f85338 | | ProjQ | 47.23 | Not recorded | 6a1e63bedbd7150506f85338 | | CALDERA | 46.85 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 46.28 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 46.1 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 45.33 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 45.19 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 44.88 | Not recorded | 6a1e63bedbd7150506f85338 | | ProjQ | 44.44 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 44.09 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 43.9 | Not recorded | 6a1e63bedbd7150506f85338 | | CALDERA | 42.5 | Not recorded | 6a1e63bedbd7150506f85338 | | ProjQ | 41.96 | Not recorded | 6a1e63bedbd7150506f85338 | | CALDERA | 41.15 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 40.98 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 40.96 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 39.55 | Not recorded | 6a1e63bedbd7150506f85338 | | ProjQ | 39.28 | Not recorded | 6a1e63bedbd7150506f85338 | | CALDERA | 38.9 | Not recorded | 6a1e63bedbd7150506f85338 | | LoftQ | 38.81 | Not recorded | 6a1e63bedbd7150506f85338 | | GPTQ+SVD-LLM | 38.38 | Not recorded | 6a1e63bedbd7150506f85338 | | AWQ+SVD-LLM | 37.82 | Not recorded | 6a1e63bedbd7150506f85338 | | AWQ+SVD-LLM | 37.51 | Not recorded | 6a1e63bedbd7150506f85338 | | AWQ+SVD-LLM | 36.67 | Not recorded | 6a1e63bedbd7150506f85338 | | AWQ+SVD-LLM | 36.61 | Not recorded | 6a1e63bedbd7150506f85338 | ### ProjQ Average Accuracy: 55.58. Result ID: 6a1e911cfbc373592c189f45. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"13B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### ProjQ Average Accuracy: 55.05. Result ID: 6a1e911cfbc373592c189f48. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"32B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### CALDERA Average Accuracy: 54.02. Result ID: 6a1e911cfbc373592c189f37. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"13B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 53.82. Result ID: 6a1e911cfbc373592c189f3d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"13B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 53.42. Result ID: 6a1e911cfbc373592c189f40. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"32B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 52.97. Result ID: 6a1e911cfbc373592c189f2b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"13B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 49.77. Result ID: 6a1e911cfbc373592c189f2e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"32B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### ProjQ Average Accuracy: 47.82. Result ID: 6a1e911cfbc373592c189f44. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### CALDERA Average Accuracy: 47.78. Result ID: 6a1e911cfbc373592c189f39. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"14B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### ProjQ Average Accuracy: 47.68. Result ID: 6a1e911cfbc373592c189f47. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"14B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### ProjQ Average Accuracy: 47.23. Result ID: 6a1e911cfbc373592c189f4b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"32B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### CALDERA Average Accuracy: 46.85. Result ID: 6a1e911cfbc373592c189f36. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 46.28. Result ID: 6a1e911cfbc373592c189f2d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"14B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 46.1. Result ID: 6a1e911cfbc373592c189f3f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"14B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 45.33. Result ID: 6a1e911cfbc373592c189f2a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 45.19. Result ID: 6a1e911cfbc373592c189f3c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 44.88. Result ID: 6a1e911cfbc373592c189f43. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"32B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### ProjQ Average Accuracy: 44.44. Result ID: 6a1e911cfbc373592c189f46. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 44.09. Result ID: 6a1e911cfbc373592c189f31. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"32B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 43.9. Result ID: 6a1e911cfbc373592c189f3e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### CALDERA Average Accuracy: 42.5. Result ID: 6a1e911cfbc373592c189f38. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### ProjQ Average Accuracy: 41.96. Result ID: 6a1e911cfbc373592c189f4a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"8B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### CALDERA Average Accuracy: 41.15. Result ID: 6a1e911cfbc373592c189f3b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"8B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 40.98. Result ID: 6a1e911cfbc373592c189f42. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"8B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 40.96. Result ID: 6a1e911cfbc373592c189f2c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 39.55. Result ID: 6a1e911cfbc373592c189f30. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"8B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### ProjQ Average Accuracy: 39.28. Result ID: 6a1e911cfbc373592c189f49. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"4B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### CALDERA Average Accuracy: 38.9. Result ID: 6a1e911cfbc373592c189f3a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"4B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### LoftQ Average Accuracy: 38.81. Result ID: 6a1e911cfbc373592c189f41. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"4B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### GPTQ+SVD-LLM Average Accuracy: 38.38. Result ID: 6a1e911cfbc373592c189f2f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Qwen3","Parameter Scale":"4B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### AWQ+SVD-LLM Average Accuracy: 37.82. Result ID: 6a1e911cfbc373592c189f33. [Reporting paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"13B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### AWQ+SVD-LLM Average Accuracy: 37.51. Result ID: 6a1e911cfbc373592c189f34. [Reporting paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### AWQ+SVD-LLM Average Accuracy: 36.67. Result ID: 6a1e911cfbc373592c189f35. [Reporting paper]() Recorded evaluation configuration: {"Model Family":"Qwen2.5-Ins","Parameter Scale":"14B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. ### AWQ+SVD-LLM Average Accuracy: 36.61. Result ID: 6a1e911cfbc373592c189f32. [Reporting paper]() Recorded evaluation configuration: {"Model Family":"LLaMA2","Parameter Scale":"7B","Quantization Bit":"2-bit","Rank (ra = rd)":"64","Zero-shot evaluation protocol":"true"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.