Hard-negative source selection for dense retrieval is usually decided only after fine-tuning and downstream evaluation. We propose $\mathrm{ECI}_{\mathrm{sem}}$, a semantic residual variant of Effective Contrastive Information (ECI) that ranks candidate negative sources using frozen target-encoder embeddings. $\mathrm{ECI}_{\mathrm{sem}}$ is training-free, not label-free: each scored example requires a query, a labeled positive, and an explicit candidate negative. $\mathrm{ECI}_{\mathrm{sem}}$ builds a weighted residual information matrix from target consistency, semantic locality, lexical residuality, and a log-determinant diversity objective. On MS MARCO negative sources, in-family $\mathrm{ECI}_{\mathrm{sem}}$ ranks LLM negatives highest among non-hybrid sources and Dense+LLM highest among hybrid sources, matching the strongest aggregate BEIR transfer results across DistilBERT, E5-base, and Contriever. Controlled ablations show that this alignment depends on using the target encoder family, while additional ablations show stability under sample-size, temperature, tokenizer, and IDF-corpus perturbations. The theory gives a local linearized link to loss reduction, while the empirical study treats downstream evaluation as the final test.

翻译：密集检索中难负样本源的选择通常仅在微调和下游评估后才能确定。我们提出$\mathrm{ECI}_{\mathrm{sem}}$，一种有效对比信息（ECI）的语义残差变体，利用冻结的目标编码器嵌入对候选负样本源进行排序。$\mathrm{ECI}_{\mathrm{sem}}$无需训练，但非无标签：每个评分示例需要一个查询、一个标注正样本和一个显式候选负样本。$\mathrm{ECI}_{\mathrm{sem}}$通过目标一致性、语义局部性、词汇残差性以及对数行列式多样性目标构建加权残差信息矩阵。在MS MARCO负样本源上，族内$\mathrm{ECI}_{\mathrm{sem}}$将LLM负样本在非混合源中排名最高，将Dense+LLM在混合源中排名最高，这与DistilBERT、E5-base和Contriever上最强聚合BEIR迁移结果一致。受控消融实验表明，这种对齐依赖于使用目标编码器族，而额外消融实验显示其在样本量、温度、分词器和IDF语料库扰动下保持稳定性。理论部分给出了与损失降低的局部线性化关联，而实证研究将下游评估视为最终测试。