预测答案引擎会引用什么：开放决策模型、重排序器与蒸馏 4B 学生在十引擎受控检索下的表现

View PDF HTML (experimental)

Abstract:Generative engine optimization needs a scorer that ranks candidate pages by the share of an answer they will receive as citations. We measure pretrained rerankers, a hosted typed decision model (Jev), eight open decision models and distilled students on GEO-Bench queries answered by ten engines that receive the same candidate sources in their prompt. Where several engines answer the same query, they agree with one another at a within-group Spearman correlation of 0.704. Jev and two open 27B decision models, used without training, are the strongest scorers at 0.575 to 0.578. A Qwen3-Reranker-4B fine-tuned with low-rank adapters on teacher labels for other GEO-Bench queries, released as GEO Citation Ranker 4B, reaches 0.542 to 0.549 across three seeds, and its differences with the 27B models and with Jev are not distinguishable from zero on 229 queries. Most of the remaining distance between the student and Jev disappears when Jev scores one passage at a time, as the student does. Students trained on teacher labels and on observed engine citations reach the same level, and labels from Jev alone match the two-teacher target at 0.6B. We release the model, the benchmark, the predictions of every reported scorer and the code.

Comments:	12 pages, 1 figure, 4 tables. Model: this https URL ; benchmark and code: this https URL
Subjects:	Information Retrieval (cs.IR); Artificial Intelligence (cs.AI)
Cite as:	arXiv:2610.38090 [cs.IR]
	(or arXiv:2610.38090v1 [cs.IR] for this version)
	https://doi.org/10.48550/arXiv.2610.38090 arXiv-issued DOI via DataCite

Submission history

From: André-Louis Rochet [view email]
[v1] Wed, 28 Oct 2026 00:11:13 UTC (40 KB)