Relevant but Incomplete: Referential Dangling as a Paradigm-Level Failure Mode in Hard Prompt Compression
- 类型：arxiv
- 标识：2608.04569
- 链接：https://arxiv.org/abs/2608.04569
- 主分类：rag
- 形态：method
- 被引：0
- 被引来源：Semantic Scholar
- S2被引：0
- 影响力被引：0
- TLDR：A compact classifier is trained to rank omitted sentences by whether they are needed to interpret retained text and reinsert the top-ranked candidates without support annotations at inference, training a compact classifier to optimize both relevance and referential completeness.
- 待LLM分类：否
- 标题中文：相关但不完整：指代悬空作为硬提示压缩中的范式级失效模式
- TLDR中文：训练一个紧凑分类器，对被省略的句子按其是否为理解保留文本所必需进行排序，并在推理时无需支撑标注地回插排名靠前的候选，同时优化相关性与指代完整性。
- 来源文件：
- /inbox/tom/_candidates/2026-08-11-rag-retrieval-reranking-candidates.json
- /inbox/tom/_candidates/2026-08-11-agent-rag-longcontext-candidates.json
- [S2 enrich]