metadata
license: apache-2.0
language:
- zh
tags:
- human-feedback
size_categories:
- n<1K
尝试解决"llm repetition problem"，使用分词模型对oaast语料进行“结巴化”数据增强，提供更强的重复内容拒绝效果。
Attempts to solve the "llm repetition problem" by using a segmentation model to enhance the oaast corpus with "stuttering" data to provide stronger rejection of duplicate content.
其次，还过滤掉了所有自我认知的微调样本。
Second, it also filters out all the fine-tuned samples of self-cognition.
files:
- oaast_rm_zh_jieba.jsonl : word level repeat
- oaast_rm_zh_sent_jieba.jsonl : sentence level repeat