This dataset contains text generation outputs from OpenAI's GPT-OSS-20B model across multiple evaluation benchmarks, with generation limited to 512 tokens.
Dataset Description
The dataset captures GPT-OSS-20B's text generation behavior when responding to prompts from established AI evaluation benchmarks. Each example includes the original prompt, the model's generated response, and token statistics.
Benchmark Coverage… See the full description on the dataset page: https://huggingface.co/datasets/AmanPriyanshu/GPT-OSS-20B-benchmark-rollouts-512-tokens.
