12 2024 档案

摘要:目的: reduce bias of LLMs(length, concreteness, empty reference, content continuation, nested instruction, familiar knowledge) Tool: OffsetBias: pairwis 阅读全文
posted @ 2024-12-30 18:58 雪溯 阅读(2) 评论(0) 推荐(0) 编辑
摘要:Abstract 本文: Speculative RAG Task: improving retrieval results by combining RAG with LLMs refinement Method: 利用large Generalist LM大点的通用模型来验证RAG drafts 阅读全文
posted @ 2024-12-25 01:54 雪溯 阅读(17) 评论(0) 推荐(0) 编辑
摘要:Abstract good words: subjectivity, variability, scale Task: Survey of LLM-as-a-Judge, benchmark & evaluation of LLM-as-a-Judge systems Core question: 阅读全文
posted @ 2024-12-21 00:46 雪溯 阅读(28) 评论(0) 推荐(0) 编辑
摘要:Abstract Task: Defense LLM from prompt injection attacks Tool: TaskTracker Methods: use activation deltas( the difference in activations before and af 阅读全文
posted @ 2024-12-13 15:58 雪溯 阅读(9) 评论(0) 推荐(0) 编辑
摘要:Abstract Github: https://github.com/JailbreakBench/jailbreakbench https://jailbreakbench.github.io/ Task: Opensource benchmark an evolving repository 阅读全文
posted @ 2024-12-10 22:42 雪溯 阅读(18) 评论(0) 推荐(0) 编辑

点击右上角即可分享
微信分享提示