snap-stanford/socsci210_evals_baseline_llama3_instruct

--- dataset_info: features: - name: index dtype: int64 - name: prompt dtype: string - name: system dtype: string - name: ground_truth dtype: string - name: scale dtype: string - name: post_id dtype: string - name: persona dtype: string - name: generation dtype: string - name: prediction dtype: string - name: accuracy dtype: float64 - name: has_tags dtype: bool - name: scale_min dtype: int64 - name: scale_max dtype: int64 splits: - name: train num_bytes: 17390137 num_examples: 10000 download_size: 3787008 dataset_size: 17390137 configs: - config_name: default data_files: - split: train path: data/train-* ---