Can LLM User-Proxies Stand In for Humans in Reliance Measurement? Replicating Rel-A.I. with Simulated Users

Shuai Wang; Rui Hu

cs.AIOct 19, 2026

Can LLM User-Proxies Stand In for Humans in Reliance Measurement? Replicating Rel-A.I. with Simulated Users

Authors: Shuai Wang, Rui Hu

Organizations: Zhejiang University

Abstract

Can language-model user proxies reproduce human reliance on an AI answer across contexts? Published simulated-user studies do not establish validity for this partial-answer decision, and Rel-A.I.'s released human record omits trial-level choices. We audit its 90 printed questions and 15 domain-by-marker reliance rates using sampled decisions and counterbalanced A/B probabilities from open models through 72B parameters. A matched crossover varies the printed agent specialty within each question, separating question content, label averages, and their interaction. Mistral Small's strong-marker math-minus-law label marginal is -0.4 percentage points, yet matching the agent label to the question domain raises reliance by 13.3 points on average across the five domains. The best A/B proxy has 22.5-point absolute-rate error and 8.3-point domain-effect error against the human table. A two-sample independent-binomial human replication has 3.5-point expected domain-effect error under the stated working model. The interaction is model dependent, and inference layout changes identical-prompt probabilities. This single-study audit does not establish proxy validity for human effect sizes or general directional screening.

Figures & tables

Figure 1: The audit pairs a published human table with matched proxy questions. The left panel shows the printed Rel-A.I. questions and epistemic-marker classes; item-to-marker assignments are reconstructed. The middle panel shows the matched question, marker, agent-label, and option-order crossover. Lollipops at right compare 15-cell rate errors for two A/B proxies with a plug-in independent human resample.

Figure 2: Sampled proxies depart from human absolute rates. Each panel places the 15 default-prompt rates of one nonconstant proxy against published human rates. The dashed diagonal is equality. Symbol shape and color identify the model; lighter symbols indicate weaker marker classes. Qwen 0.5B’s constant-zero rates remain in Table 1 .

Domain	Marker	Human	Qwen 7B	Mistral 7B	Phi 3.5
College math	Strong	90.1	5.2	9.9	94.8
College math	Moderate	58.5	9.2	3.3	94.8
College math	Weak	14.6	0.0	0.0	83.3
Abstract algebra	Strong	87.0	6.8	8.3	89.6
Abstract algebra	Moderate	59.8	10.6	4.4	86.3
Abstract algebra	Weak	12.0	0.0	0.0	66.1

Table 1: Reliance rate (%) in every published domain–marker cell. Human values are from Rel-A.I. Table 4; proxy values are default-prompt estimates on reconstructed assignments.

Proxy	Rate MAE	Rate $\rho$	Effect MAE	H hits	Joint res.	Joint hits
Qwen 0.5B	50.1	–	4.6	0	0	0
Qwen 7B	42.6	0.51	12.0	3	3	1
Mistral 7B	38.3	0.61	12.4	0	5	0
Phi 3.5	37.8	0.55	10.3	3	2	0
Oracle overall mean	27.2	–	4.6	0	–	–
Oracle marker mean	3.2	0.94	4.6	0	–	–

Table 2: Default-prompt agreement across all 15 rates and 30 domain-pair contrasts. Errors are percentage points. H hits are matched human-resolved pairs with a proxy effect of at least 1 pp. Joint res. counts pairs for which both working intervals exclude zero; joint hits counts their matching directions. Human resample is an independent-binomial plug-in reference; oracle rules use the scored human table.

Figure 3: The sampled mathematics-minus-law effects remain uncertain. Colored points show default-prompt proxy estimates and vertical bars show percentile intervals from resampling prompt replicates and questions. Grey crosses are published human effects. The human working intervals are evaluated separately because participant records are unavailable.

Domain	Marker	Human	Qwen 7B	Qwen 32B	Qwen 72B	Mistral Small
College math	S	90.1	37.1	74.0	55.2	43.2
College math	M	58.5	18.2	43.3	48.8	35.0
College math	W	14.6	1.4	0.0	0.9	13.5
Abstract algebra	S	87.0	21.7	59.2	62.8	36.4
Abstract algebra	M	59.8	16.1	15.6	51.3	27.8
Abstract algebra	W	12.0	2.9	0.0	0.6	11.0

Table 3: Human and counterbalanced A/B reliance rates (%). Every value is a point estimate. S, M, W denote strong, moderate, and weak epistemic markers.

Protocol	Rate MAE	Rate $\rho$	Domain MAE	Marker MAE	$\geq$ 1 pp	H-res. hits
Oracle overall	27.2	–	4.6	77.3	0	0
Oracle marker	3.2	0.94	4.6	2.2	0	0
Posthoc STEM	2.8	0.98	4.0	2.2	17	8
Qwen 7B	43.5	0.40	9.1	66.5	18	7
Qwen 32B	35.5	0.88	20.0	45.2	12	5
Qwen 72B AWQ	30.9	0.87	19.6	42.5	15	5

Table 4: A/B control, human resampling reference, and table-fitted diagnostic rules. Errors are percentage points; marker error scores five strong-minus-weak contrasts. The 1-pp sign count requires both human and proxy effects to reach that magnitude. H-res. hits counts matched human-resolved pairs out of eight. The human row averages independent-binomial samples using published rates as plug-in probabilities, so its counts can be fractional. Oracle and posthoc rules are descriptive, not deployable predictors.

Proxy	Diag.	Content	Label	Cong. S	Cong. M	Cong. W
Qwen 7B	9.6	5.2	3.4	6.9	4.2	0.3
Qwen 72B AWQ	19.7	10.2	5.6	23.7	15.0	0.2
Mistral Small	8.3	6.1	5.3	13.3	9.6	2.4
Zero domain effect	4.6	4.6	4.6	–	–	–
Two-sample human	3.5	–	–	–	–	–

Table 5: Matched crossover point estimates (pp). Domain MAE compares the human 30-pair table with diagonal, content, or label marginals. Congruence is reliance under the matching agent label minus the mean of four other labels, averaged over the five source-question domains. Human and zero-effect references apply only to the domain-MAE columns.

Figure 4: The matching-label effect persists when the global label marginal is small. Strong-marker matching-label lift compares each question’s own agent-domain label with the mean of its four other labels, within source domain. Points are measured domain means from the matched crossover; zero denotes no within-domain label preference.

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

Human model	Pairs	STEM	Qwen 7B	Qwen 32B	Qwen 72B	Mistral Small
Discrete B	8	8	7	5	5	4
Corr. 0	10	10	9	6	6	5
Corr. 0.3	14	13	12	8	9	7

Appendix

Table 6: Matched human-resolved directions under three working interval choices. B uses the discrete binomial bootstrap; zero and 0.3 are within-person correlation settings in the normal sensitivity model. The STEM rule is fitted to the human table. Each entry is a matched-pair count; all proxy effects must reach 1 pp.

Marker	Pair	Human	B	0	.3	Phi	Q7	Q32	Q72	Mistral
S	Math–Alg.	3.1	–	–	–	5.2	15.4	14.8	-7.6	6.8
S	Math–Phil.	1.0	–	–	–	-2.6	37.1	73.9	42.1	-9.4
S	Math–Rel.	6.2	–	–	Y	-4.7	36.6	52.2	12.5	-5.9
S	Math–Law	6.8	–	Y	Y	9.0	37.1	68.5	53.4	9.8
S	Alg.–Phil.	-2.1	–	–	–	-7.8	21.7	59.1	49.7	-16.1
S	Alg.–Rel.	3.1	–	–	–	-9.9	21.3	37.4	20.1	-12.6

Appendix

Table 7: Within-marker domain effects in percentage points. B denotes the discrete-binomial working interval; 0 and .3 denote normal within-person-correlation sensitivity settings. Y marks a human effect whose interval excludes zero. Proxy values are point estimates; full intervals remain with experiment details.

Domain	Marker	Qwen 7B	Mistral 7B	Phi 3.5
College math	S	4.2	8.3	80.2
College math	M	5.6	5.0	77.1
College math	W	0.0	0.5	44.3
Abstract algebra	S	1.0	9.4	66.1
Abstract algebra	M	3.1	5.2	65.2
Abstract algebra	W	0.0	2.1	20.8

Appendix

Table 8: Diverse-persona sampled reliance rates (%) for every domain and marker class. Point estimates are shown; the corresponding default rates appear in Table 1 .

Proxy	Questions	Math	Algebra	Philosophy	Religion	Law
Qwen 7B	Math	39.0	19.6	22.0	12.1	9.9
	Algebra	32.4	21.8	5.4	3.0	1.3
	Philosophy	0.0	0.0	0.0	0.0	0.0
	Religion	0.1	0.0	0.2	0.4	0.0
	Law	0.1	0.0	0.2	0.0	0.0
Qwen 72B AWQ	Math	55.4	44.0	16.8	8.2	6.7

Appendix

Table 9: Strong-marker crossover matrix: source question domain by printed agent-domain label. Every cell is the balanced reliance rate (%). Matching labels are on the diagonal.

Figure 5: Content and label marginals differ across models. Mathematics-minus-law effects are shown for strong (S), moderate (M), and weak (W) markers. Blue circles compare source question sets after averaging over agent labels; vermilion squares compare agent labels after averaging over all questions. Grey diamonds are published human diagonal effects and are not matched label-only effects. Vertical bars are percentile intervals from resampling questions within source domain.

User Simulation

User simulation models human behavior and interaction to test, train, and evaluate AI systems with synthetic users.

See all 128 papers

Language Models Give Group-Linked Fields the Wrong Weight
Aug 24, 2026Zengqing Wu, Chuan XiaoLanguage Model Bias Evaluation Demographic Bias in Language Models
University of Osaka Osaka, Japan
Understanding Reliability in LLM-based Human Behavior Simulation
Sep 16, 2026Pei Wang, Lei Wang, Yuanzi Li +1Human Behavior Simulation Large Language Model-Based Social Simulation
Gaoling School of Artificial Intelligence, Renmin University of China
Beyond Averages: Evaluating LLMs on Human Survey Replication at the Distributional Level
Jun 8, 2026Jeonghyeon Moon, Jiwon Kim, Yeheum Lah +2LLM Evaluation Human Behavior Simulation
Ewha Womans University

Aug 24, 2026J/K move · Enter open · S save

Language Models Give Group-Linked Fields the Wrong Weight

Zengqing Wu, Chuan Xiao

University of Osaka Osaka, Japan

Language Model Bias Evaluation Demographic Bias in Language Models Causal Interventions in Language Models Clinical Language Model Evaluation

Open paper AskPDF arXiv

Large language models are increasingly used to judge people, and their answers can depend on inputs correlated with a protected attribute such as race, even when that attribute is never shown. Interpreting this reliance requires knowing how much predictive information those inputs add, but audits often report only whether a decision changes. We introduce an evidence-referenced audit that compares proxy reliance with two references, the known outcome rule and a Bayesian linear regression fitted to the same labeled examples, while measuring ranking accuracy separately. We apply it to four models in a simulated patient-ranking task with known field weights. We find four results. First, at zero predictive value, the models still relied on fields that carried no information, while the same-example learner did not. Second, as proxies became predictive, reliance grew far less than the reference. Third, display format mattered. Ordered labels drew more reliance than arbitrary labels, a direction that held in all four models. Fourth, ranking accuracy did not consistently track these changes in reliance. Based on these findings, we argue that decision changes and accuracy cannot show whether proxy reliance is warranted by the evidence. Audits should report reliance against an explicit predictive reference, alongside accuracy. Ordered fields whose values move with group membership, such as scores or income bands, are a useful starting point. Together, these results make it possible to measure how far reliance departs from the evidence, rather than infer it from observed choices alone.

Abstract

User Simulation

Explore similar work

Language Models Give Group-Linked Fields the Wrong Weight

Understanding Reliability in LLM-based Human Behavior Simulation

Beyond Averages: Evaluating LLMs on Human Survey Replication at the Distributional Level