The Librarian Who Refused to Code: Model-Dependent Identity Enactment in LLM Code Generation

A pre-registered study on arXiv (2607.17420v1) found that biographical personas in system prompts produce model-dependent effects on LLM code generation, with the research-librarian persona causing Claude Opus to issue in-character disclaimers in 55 of 60 responses and 12 genuine no-code responses, lowering mean correctness from 0.92 to 0.67, while GPT-5.5 showed no such behavior. The minimalist engineer persona reduced Claude Opus's visible output by 30% without improving correctness, and the thorough engineer persona increased output without a correctness gain, leading the researchers to conclude that personas act as model-dependent behavioral-policy biases rather than universal quality interventions.

arXiv:2607.17420v1 Announce Type: new Abstract: Biographical personas are widely used in system prompts, but their effects on code generation are rarely evaluated under controlled, pre-registered conditions. We tested four prompt conditions no persona, two engineer personas, and a research-librarian persona , 12 code-generation tasks, two frontier models, and five runs per cell 480 completions . Persona effects differed between the two tested models. Under the pre-registered mixed-effects analysis, the condition-by-model interaction was significant for provider-reported output tokens; a post-hoc visible-character measure showed the same qualitative pattern. Six GPT-5.5 completions were length-capped and are reported separately. On Claude Opus, the minimalist engineer persona reduced visible output by 30% 33% in provider tokens without improving correctness, while the thorough engineer persona increased output without a correctness gain. In an exploratory post-hoc analysis, the librarian persona elicited in-character disclaimers in 55 of 60 Opus responses and 12 genuine no-code responses, lowering mean correctness from 0.92 to 0.67. GPT-5.5 produced neither behavior in its 59 non-truncated responses. These results are consistent with personas acting as Model-Dependent behavioral-policy biases rather than universal quality interventions. We release raw completions, derived scores, analysis artifacts, a pre-registration document, and an execution gate log; end-to-end test-based rescoring requires an unreleased task harness.