datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
CLIP-CC
📚 CLIP-CC Dataset (Movie Clips Edition)
Paper | arXiv | Project Page | Benchmark Code (CLIP-CC-Bench) | Dataset Repo (CLIP-CC)
CLIP-CC is a curated dataset for long-form video description: 200 movie clips sourced from YouTube, each about 90 seconds long (~5 hours in total) and drawn from more than 140 films spanning 1959–2024, each paired with one human-written English reference description averaging 402 ± 208 words. The references were written by four graduate-student… See the full description on the dataset page: https://huggingface.co/datasets/MINT-SDSU/CLIP-CC.sdsdsdsdsOpenR1-SDS-Base-Generations-seed303
OpenR1-SDS-Base-Generations-seed303
Raw code generations from the untrained base model (Qwen2.5-Coder-14B-Instruct)
evaluated on the Synergistic Dependency Selection (SDS) benchmark.
Contents
generations.jsonl: 64,000 lines (64 samples x 1,000 problems)
Seed: 303
Model: Qwen/Qwen2.5-Coder-14B-Instruct (no RL fine-tuning)
Decoding: temperature=0.6, 64 independent samples per problem
System prompt: Identical to Hero (from config_hero.yaml)
Purpose
These… See the full description on the dataset page: https://huggingface.co/datasets/IDEALLab/OpenR1-SDS-Base-Generations-seed303.Opus-4.6-Reasoning-3000x-filteredFiltered from: https://huggingface.co/datasets/crownelius/Opus-4.6-Reasoning-3000x
The original dataset has 979 refusals, I removed these in this version.
RAFT_SDSnewtSDS_3MOpenR1-SDS-Base-Generations-seed202
OpenR1-SDS-Base-Generations-seed202
Raw code generations from the untrained base model (Qwen2.5-Coder-14B-Instruct)
evaluated on the Synergistic Dependency Selection (SDS) benchmark.
Contents
generations.jsonl: 64,000 lines (64 samples x 1,000 problems)
Seed: 202
Model: Qwen/Qwen2.5-Coder-14B-Instruct (no RL fine-tuning)
Decoding: temperature=0.6, 64 independent samples per problem
System prompt: Identical to Hero (from config_hero.yaml)
Purpose
These… See the full description on the dataset page: https://huggingface.co/datasets/IDEALLab/OpenR1-SDS-Base-Generations-seed202.OpenR1-SDS-Base-Generations-seed101
OpenR1-SDS-Base-Generations-seed101
Raw code generations from the untrained base model (Qwen2.5-Coder-14B-Instruct)
evaluated on the Synergistic Dependency Selection (SDS) benchmark.
Contents
generations.jsonl: 64,000 lines (64 samples x 1,000 problems)
Seed: 101
Model: Qwen/Qwen2.5-Coder-14B-Instruct (no RL fine-tuning)
Decoding: temperature=0.6, 64 independent samples per problem
System prompt: Identical to Hero (from config_hero.yaml)
Purpose
These… See the full description on the dataset page: https://huggingface.co/datasets/IDEALLab/OpenR1-SDS-Base-Generations-seed101.