Hai-Tao Zheng, Gang Wang, Jieming Zhu, Xiaopeng Li, Zhicheng Zhang, Kuicai Dong, Jingyu Li, Jiwei Tang, Zhaocheng Du, Qianhui Zhu, Fengyuan Lu, Wang Jiaheng
We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. If it links a repository, it is listed below.
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Hard negative mining has become the dominant strategy for training retrievers, yet it faces intrinsic limitations: negatives are bounded by corpus availability, selected by retriever score rather than diagnostic value, and increasingly contaminated by false positives as the retriever improves. LLM-based synthesis offers a principled alternative, where negatives that are unconstrained, targeted, and free from false positive risk. But we show that naïvely incorporating generated negatives into contrastive learning often degrades retrieval performance. We identify and formalize the root cause as a generative-discriminative gap: LLM generation optimizes for fluent, plausible text, while contrastive learning demands strategic violations of relevance at the decision boundary. Our analysis reveals two compounding failure modes: discriminative-agnostic generation, where the LLM lacks an explicit model of query information needs and defaults to generic or topic-drifted text that provides no contrastive signal; and source-dependent shortcuts, where distributional artifacts enable the model to distinguish negatives by origin
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2606.01304")
get_code_for_paper("2606.01304")
have("2606.01304")
Connect an agent — have() is free.