JingweiNi/train_prm800k_gpt-oss-120b_annotated_qwen3_1.7b_thinking_5000_shards_4_7_16

该数据集包含多个字段，如question（问题）、answer（答案）、input_ids（输入ID）、reply（回复）、original_index（原始索引）、claims（声明）和verified（已验证）。claims字段进一步细分为aligned_token_ids（对齐的标记ID）、claim_text（声明文本）和sentence（句子）。数据集只有一个名为train的分割，包含2,500个示例，总大小为44,325,911字节。该数据集适用于涉及问答和声明验证的任务。

The dataset includes various fields such as question, answer, input_ids, reply, original_index, claims, and verified. The claims field is further subdivided into aligned_token_ids, claim_text, and sentence. The dataset has a single split named train with 2,500 examples and a total size of 44,325,911 bytes. This dataset is designed for tasks involving question-answering and claim verification.