How model-side context position shapes generation opportunity—and why it must not be confused with citation prominence.
Position is one of the more reproducible factors
Martinez’s review identifies query-document relevance and context position among the most consistently supported determinants of source use in controlled research. Sources placed earlier or more prominently in context often receive greater opportunity to be cited or used.
This does not make position the only mechanism. Strong semantic relevance can outweigh a weak positional advantage.
Model position and user-visible position belong to different stages
Context rank describes where material appears in the input available to the generator. Citation or answer position describes what the user sees in the final response.
Context position affects generation opportunity. Citation position affects observable prominence.
A source can be early in context yet cited late—or visible early without researchers knowing its hidden context order.
Treat context rank as its own dependent variable
When a reproducible system exposes reranking, useful measures include top-k inclusion rate, mean context rank, recall at k, reranking score, selected passage count, allocated tokens, and context share.
Keeping these measures separate prevents context-selection effects from being hidden inside downstream citation outcomes.
Questions about this topic
Why can context position affect citation?+
Controlled studies suggest models do not use all positions equally, so highly placed relevant evidence may receive more opportunity to influence generation.
Is context position the same as first citation position?+
No. Context position is hidden model input order; citation position is the user-visible location of attribution in the answer.
Is position more important than relevance?+
Not universally. Both matter, and strong relevance can dominate a modest positional advantage.
References
Sources are listed in APA 7 style. Preprints are identified as such and should not be treated as peer-reviewed findings unless separately published.
- Martinez, O. (2026). Optimizing visibility in generative engines: A critical survey of generative engine optimization (2023–2026) [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2607.14035
- Aggarwal, P., Murahari, V., Rajpurohit, T., Kalyan, A., & Narasimhan, K. (2024). GEO: Generative engine optimization. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (pp. 5–16). Association for Computing Machinery. https://doi.org/10.1145/3637528.3671900
- Zhang, K., He, X., & Yao, J. (2026). From citation selection to citation absorption: A measurement framework for generative engine optimization across AI search platforms [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2604.25707