skbose/grokforge-eval-results

--- dataset_info: features: - name: log dtype: string - name: pattern_name dtype: string - name: reference dtype: string - name: predicted dtype: string - name: exact_match dtype: bool - name: functional_match dtype: bool - name: latency_s dtype: float64 - name: error dtype: string splits: - name: test_claude_haiku_4_5 num_bytes: 84782 num_examples: 200 - name: test_claude_opus_4_6 num_bytes: 85048 num_examples: 200 - name: test_claude_sonnet_4_6 num_bytes: 84374 num_examples: 200 download_size: 95575 dataset_size: 254204 configs: - config_name: default data_files: - split: test_claude_haiku_4_5 path: data/test_claude_haiku_4_5-* - split: test_claude_opus_4_6 path: data/test_claude_opus_4_6-* - split: test_claude_sonnet_4_6 path: data/test_claude_sonnet_4_6-* ---