Identity or Prompt Noise? A Calibrated Invariance Audit of LLM Code Generation
Source
Maksim E. Eren, Ryan Barron, Eric Michalak, Charles Nicholas, Manish Bhattarai
Author
Maksim E. Eren, Ryan Barron, Eric Michalak, Charles Nicholas, Manish Bhattarai
Date
Key takeaways · AI-distilled
The audit covers 30.73 million executed Python generations from seven checkpoints on HumanEval+ and MBPP+, using within-task randomization and false-discovery-rate control to separate persona effects from sampling and wording noise.
Occupation personas pushed CodeBLEU dispersion past its null in 10 of 14 model-benchmarkA standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.Full definition → cells, but the median excess was only 0.141 points, and pass-rate dispersion was significant in none of the six high-pass-rate cells.
Country personas led raw dispersion in 12 of 14 cells yet had a median calibrationHow well a model's confidence matches reality — a calibrated model saying "90% sure" is right about 90% of the time.Full definition → ratio of 1.00, suggesting raw counterfactual comparisons can overstate identity effects.
The authors say gender effects could not be separated from persona wording in this design, and conclude the evidence supports small occupation-linked changes in code form, not stable disadvantage to named identities.
Terms in this piece · Glossary
LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
calibration — How well a model's confidence matches reality — a calibrated model saying "90% sure" is right about 90% of the time.
Why it matters
If your coding AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → or assistant conditions on user-provided persona or profile information, occupation cues specifically and measurably shift code style and complexity, worth checking for in persona-aware coding tools.
articleWhen Who You Are Can Change the Code You Get: A Study of Persona-Induced Bias in LLM Code GenerationAnubhav Gupta, Mayara Costa Figueiredo, Leticia Santos Machado, Tanner Wright, Ivan Beschastnikh, Cleidson R. B. de Souza, Gema Rodr\'iguez-P\'erez
articleDo Influence Tactics Matter? Investigating Prompt Framing Effects in LLM Code GenerationAlex Deaconu, Anubhav Gupta, Manaal Basha, Nicholas Haydu, Gema Rodr\'iguez-P\'erez
articleA Study on the Impact of Natural Language Differences in Prompts on Automatic Code Generation Using LLMsHaruka Tokumasu, Masanari Kondo, Alexander Serebrenik, Dong Wang, Kei Koyanagi, Kotaro Noguchi, Naoyasu Ubayashi, Yasutaka Kamei