Demographic Prompting at Scale: When More Attributes Hurt LLM--Human Agreement

arXiv cs.CL Papers
demographic-prompting llm-alignment human-annotations attribute-signals neuron-probing over-specification

Summary

This paper investigates how adding demographic attributes in prompts affects LLM-human agreement across tasks, finding that while a few high-signal attributes improve alignment, over-specification degrades it. The study uses five open-source LLMs and neuron probing to show that attribute signal quality and coherence matter more than quantity.

arXiv:2607.10590v1 Announce Type: new Abstract: We investigate how annotator demographic attributes, supplied as prompt cues, shape the alignment between large language model (LLM) predictions and human annotations across five tasks. Using five open-source LLMs, we systematically vary the number and composition of demographic components in the prompt, spanning every combination from single-attribute through full-attribute configurations. Our experiments reveal three principal findings. First, alignment consistently peaks with one to three high-signal attributes and degrades under the full attribute set, establishing a clear over-specification threshold. Second, the overall magnitude of demographic influence on human annotations does not predict which attributes improve LLM alignment; instead, both the learnability and the directional coherence of each attribute's annotation signal need to be considered jointly. Third, neuron probing reveals that specialized activation correlates with alignment gains only under coherent annotation signals, and that activation volume alone does not imply steerability. Together, these results demonstrate that demographic prompting is not a monolithic intervention: its utility is highly context-dependent, shaped by attribute signal quality, task characteristics, and model architecture.
Original Article
View Cached Full Text

Cached at: 07/14/26, 04:22 AM 

# Demographic Prompting at Scale: When More Attributes Hurt LLM--Human Agreement
Source: [https://arxiv.org/abs/2607.10590](https://arxiv.org/abs/2607.10590)
[View PDF](https://arxiv.org/pdf/2607.10590)

> Abstract:We investigate how annotator demographic attributes, supplied as prompt cues, shape the alignment between large language model \(LLM\) predictions and human annotations across five tasks\. Using five open\-source LLMs, we systematically vary the number and composition of demographic components in the prompt, spanning every combination from single\-attribute through full\-attribute configurations\. Our experiments reveal three principal findings\. First, alignment consistently peaks with one to three high\-signal attributes and degrades under the full attribute set, establishing a clear over\-specification threshold\. Second, the overall magnitude of demographic influence on human annotations does not predict which attributes improve LLM alignment; instead, both the learnability and the directional coherence of each attribute's annotation signal need to be considered jointly\. Third, neuron probing reveals that specialized activation correlates with alignment gains only under coherent annotation signals, and that activation volume alone does not imply steerability\. Together, these results demonstrate that demographic prompting is not a monolithic intervention: its utility is highly context\-dependent, shaped by attribute signal quality, task characteristics, and model architecture\.

## Submission history

From: Mahammed Kamruzzaman \[[view email](https://arxiv.org/show-email/1db305f2/2607.10590)\] **\[v1\]**Sun, 12 Jul 2026 06:00:51 UTC \(9,109 KB\)

Similar Articles

LLMs Can Better Capture Human Judgments--With the Right Prompts

arXiv cs.CL

This paper presents simple prompting strategies that help large language models better capture the full distribution of human judgments, improving alignment on moral scenarios and beliefs. The authors show that asking models to report standard deviations and response proportions, along with ensuring scenario clarity, yields better agreement with human responses.

Evaluating LLMs as Human Surrogates in Controlled Experiments

arXiv cs.CL

This paper evaluates whether off-the-shelf LLMs can reliably simulate human responses in controlled behavioral experiments by comparing LLM-generated data with human survey responses on accuracy perception. The findings show that while LLMs capture directional effects and aggregate belief-updating patterns, they do not consistently match human-scale effect magnitudes, clarifying when synthetic LLM data can serve as behavioral proxies.

Re-Centering Humans in LLM Personalization

arXiv cs.CL

This paper studies the gap between synthetic and human data for evaluating LLM personalization across three stages: attribute extraction, relevance matching, and response generation. Results show models perform worse on real human data, and the authors introduce lightweight training interventions to improve alignment.

Review Arcade: On the Human Alignment and Gameability of LLM Reviews

Hugging Face Daily Papers

This paper investigates the alignment of LLM-generated reviews with human judgment using 1k real ACL 2025 submissions, finding limited agreement, instability across models/prompts, and a method to artificially inflate scores without meaningful changes. The authors advise against relying solely on LLM reviews and call for discussion on their use in handling increasing submission volumes.

The Geometry of LLM-as-Judge: Why Inter-LLM Consensus Is Not Human Alignment

arXiv cs.CL

This paper geometrically analyzes why LLMs acting as judges agree strongly with each other but weakly with humans, finding that inter-LLM consensus reflects a collapsed subspace rather than true human alignment on subjective rubrics. Post-hoc calibration on human data improves alignment, but even calibrated LLMs fall short of human reliability.