diff --git a/playground/dataset_variants/distill_filter.py b/playground/dataset_variants/distill_filter.py index a94e704cf..ccb104657 100644 --- a/playground/dataset_variants/distill_filter.py +++ b/playground/dataset_variants/distill_filter.py @@ -9,7 +9,7 @@ def gen_distill_data(input_file, output_file): distill_data = [] for conv in all_data: - if conv["model"] in ["claude-1", "claude-instant-1", "gpt-4", "claude-2"]: + if conv["model"] in ["claude-1", "claude-instant-1", "claude-2", "gpt-4", "gpt-3.5-turbo"]: distill_data.append(conv) print("distill data size", len(distill_data)) diff --git a/playground/dataset_variants/upvote_filter.py b/playground/dataset_variants/upvote_filter.py index f8d6d630a..6744d1423 100644 --- a/playground/dataset_variants/upvote_filter.py +++ b/playground/dataset_variants/upvote_filter.py @@ -10,7 +10,7 @@ def gen_upvote_data(input_file, output_file): upvote_data = [] for conv in all_data: - if conv["model"] not in ["claude-1", "claude-instant-1", "gpt-4", "claude-2", "palm-2", "gpt-3.5-turbo"]: + if conv["model"] not in ["claude-1", "claude-instant-1", "claude-2", "palm-2", "gpt-4", "gpt-3.5-turbo"]: upvote_data.append(conv) print("upvote data size", len(upvote_data))