BehaviorBench: Benchmarking Foundation Models for Behavioral Science Tasks
Quick Answer
This paper shows that BehaviorBench benchmarks foundation models for behavioral science, revealing that Be.FM-1.5 excels in distributional alignment while maintaining competitive individual-level metrics.
Quick Take
Proprietary models perform well in individual tasks but lack broader population alignment, highlighting the need for behavioral adaptation in AI systems.
Key Points
- BehaviorBench evaluates models on behavior prediction, decision-making, trait inference, and knowledge application.
- Be.FM-1.5 shows superior distributional alignment compared to general-purpose models.
- Proprietary models excel in individual-level predictions but lack population-level validity.
- BehaviorBench serves as a foundation for developing behaviorally aligned AI systems.
- Models can be accessed at https://umich-foreseer.github.io/behaviorbench/.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Foundation models have been increasingly applied to behavioral science domains such as psychology, sociology, and economics. While these models show promise in individual tasks such as survey response prediction and human-subject experiment simulation, there remains no systematic understanding of how well they perform across diverse behavioral science tasks, contexts, and populations. We introduce BehaviorBench, a comprehensive benchmark that evaluates foundation models along four core capabilities: (1) behavior prediction and simulation, (2) strategic decision-making, (3) subject-trait inference, and (4) behavioral knowledge application. Crucially, BehaviorBench evaluates model outputs at both the individual and distributional levels, capturing not only per-subject accuracy but also population-level alignment, an essential requirement for behavioral validity. Leveraging the tasks in BehaviorBench, we further develop this http URL-1.5, extending the this http URL family of behavioral foundation models fine-tuned on behavioral data. Our results reveal a considerable gap: proprietary general-purpose models excel at individual-level prediction and knowledge-intensive tasks, whereas behavioral foundation models, fine-tuned on behavioral data, achieve substantially stronger distributional alignment. Notably, this http URL-1.5 leads on distributional metrics and remains competitive on individual-level metrics, suggesting that proper behavioral adaptation can close the gap. Our results highlight the importance of distributional evaluation, establish BehaviorBench as a foundation for developing and assessing behaviorally aligned AI systems, and demonstrate this http URL-1.5's potential for a broad range of behavioral science studies. Our BehaviorBench and this http URL-1.5 models can be accessed via this https URL.
| Subjects: | Computation and Language (cs.CL); Machine Learning (cs.LG) |
| Cite as: | arXiv:2606.24162 [cs.CL] |
| (or arXiv:2606.24162v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2606.24162 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jin Huang [view email]
[v1]
Tue, 23 Jun 2026 05:30:54 UTC (1,726 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CL
See more →TriAgent: Divergence-Aware Committees for Cost-Efficient Financial Sentiment Analysis
TriAgent introduces a cost-efficient multi-agent system for financial sentiment analysis, combining VADER, FinBERT, and Qwen2.5. It achieves an F1 score of ~0.87 with significant savings of $9.3M/year at a 10M-user scale compared to GPT-4o-mini, while also detecting hallucinations with an AUC of 0.90.