[BFCL] Dataset Revamp Initiative #765
Description
Describe the issue
There has been constant feedback for dataset ground truth inconsistency and our team is tasked on a 2-week initiative to re-scrutinize across V3 dataset issues with several objectives:
- Eliminate Ground Truth mismatches against user questions.
- Polish ambiguous prompts that have unclear user intents to eliminate biased-judgement and saturation.
Proposed Change Tacker
- live_simple: [BFCL Dataset Revamp 2/n] Live Dataset Fix (Simple, Parallel, Parallel Multiple) #737
- live_multiple: [BFCL Dataset Revamp 3/n] Live Dataset Fix (Multiple) #739
- live_parallel: [BFCL Dataset Revamp 2/n] Live Dataset Fix (Simple, Parallel, Parallel Multiple) #737
- live_parallel_multiple: [BFCL Dataset Revamp 2/n] Live Dataset Fix (Simple, Parallel, Parallel Multiple) #737
- live_irrelevance: [BFCL Dataset Revamp 4/n] Live Irrelevance #763
- live_relevance: [BFCL Dataset Revamp 4/n] Live Irrelevance #763
- multi_turn_base: [BFCL Dataset Revamp 1/n] Multi-Turn (Part 1) #740
- multi_turn_miss_func:
- multi_turn_miss_param:
- multi_turn_long_context: