# General Knowledge on Knowledge Benchmarks (ARC-C, ARC-E, MMLU, GPQA) (test) Source: [General Knowledge on Knowledge Benchmarks (ARC-C, ARC-E, MMLU, GPQA) (test)]() Collection index: [llms.txt]() Data updated: 2026-02-27T10:52:01.256Z Evaluation of model performance across four key general knowledge benchmarks: ARC Challenge (ARC-C), ARC Easy (ARC-E), MMLU, and GPQA, reported on the test split. 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 | General Knowledge | | Dataset (including any stated split) | Knowledge Benchmarks (ARC-C, ARC-E, MMLU, GPQA) (test) | | Metric | ARC-C | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | 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 - [ARC-C](): Higher is better; unit: PERCENT. - [ARC-E](): Higher is better; unit: PERCENT. - [MMLU](): Higher is better; unit: PERCENT. - [GPQA](): Higher is better; unit: PERCENT. - [Average Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | ARC-C | Result date | Report paper | | --- | --- | --- | --- | | Task Arithmetic | 83.05 | Not recorded | 69647fe953fdd5c36174478b | | Linear | 80.68 | Not recorded | 69647fe953fdd5c36174478b | | LED | 80.34 | Not recorded | 69647fe953fdd5c36174478b | | FuseLLM | 80 | Not recorded | 69647fe953fdd5c36174478b | | Finance (WiroAI-Finance-Llama-8B) | 79.66 | Not recorded | 69647fe953fdd5c36174478b | | ReasonAny | 79.32 | Not recorded | 69647fe953fdd5c36174478b | | DARE | 77.29 | Not recorded | 69647fe953fdd5c36174478b | | TIES | 73.9 | Not recorded | 69647fe953fdd5c36174478b | | Reasoning (DeepSeek-R1-Distill-Llama-8B) | 70.85 | Not recorded | 69647fe953fdd5c36174478b | | Reasoning | 64.75 | Not recorded | 69647fe953fdd5c36174478b | | ReasonAny | 64.31 | Not recorded | 69647fe953fdd5c36174478b | | Task Arithmetic | 63.56 | Not recorded | 69647fe953fdd5c36174478b | | Biomedicine | 60.34 | Not recorded | 69647fe953fdd5c36174478b | | Linear | 60.34 | Not recorded | 69647fe953fdd5c36174478b | | LED | 32.54 | Not recorded | 69647fe953fdd5c36174478b | | TIES | 21.36 | Not recorded | 69647fe953fdd5c36174478b | | DARE | 17.97 | Not recorded | 69647fe953fdd5c36174478b | | FuseLLM | 0 | Not recorded | 69647fe953fdd5c36174478b | ### Task Arithmetic ARC-C: 83.05. Result ID: 69a1774c35dfcabb00e73911. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging":"Task Arithmetic","Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### Linear ARC-C: 80.68. Result ID: 69a1774c35dfcabb00e73910. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Merging":"Linear","Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### LED ARC-C: 80.34. Result ID: 69a1774c35dfcabb00e73915. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging":"LED","Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### FuseLLM ARC-C: 80. Result ID: 69a1774c35dfcabb00e73914. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging":"FuseLLM","Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### Finance (WiroAI-Finance-Llama-8B) ARC-C: 79.66. Result ID: 69a1774c35dfcabb00e7390e. [Reporting paper]() Recorded evaluation configuration: {"Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### ReasonAny ARC-C: 79.32. Result ID: 69a1774c35dfcabb00e73916. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging":"ReasonAny","Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### DARE ARC-C: 77.29. Result ID: 69a1774c35dfcabb00e73913. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging":"DARE","Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### TIES ARC-C: 73.9. Result ID: 69a1774c35dfcabb00e73912. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Merging":"TIES","Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### Reasoning (DeepSeek-R1-Distill-Llama-8B) ARC-C: 70.85. Result ID: 69a1774c35dfcabb00e7390f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Family":"Llama3.1-8B"}. This record does not assert independent reproduction. ### Reasoning ARC-C: 64.75. Result ID: 69a1774c35dfcabb00e73717. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Domain Expert (DeepSeek-R1-Distill-Qwen-7B)"}. This record does not assert independent reproduction. ### ReasonAny ARC-C: 64.31. Result ID: 69a1774c35dfcabb00e7371e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Class":"Model Merging"}. This record does not assert independent reproduction. ### Task Arithmetic ARC-C: 63.56. Result ID: 69a1774c35dfcabb00e73719. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Class":"Model Merging"}. This record does not assert independent reproduction. ### Biomedicine ARC-C: 60.34. Result ID: 69a1774c35dfcabb00e73716. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Domain Expert (Meditron3-Qwen2.5-7B)"}. This record does not assert independent reproduction. ### Linear ARC-C: 60.34. Result ID: 69a1774c35dfcabb00e73718. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Method Class":"Model Merging"}. This record does not assert independent reproduction. ### LED ARC-C: 32.54. Result ID: 69a1774c35dfcabb00e7371d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Class":"Model Merging"}. This record does not assert independent reproduction. ### TIES ARC-C: 21.36. Result ID: 69a1774c35dfcabb00e7371a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Class":"Model Merging"}. This record does not assert independent reproduction. ### DARE ARC-C: 17.97. Result ID: 69a1774c35dfcabb00e7371b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Class":"Model Merging"}. This record does not assert independent reproduction. ### FuseLLM ARC-C: 0. Result ID: 69a1774c35dfcabb00e7371c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method Class":"Model Merging"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.