Rigidity in LLM Bandits with Implications for Human-AI Dyads
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911498117316608 |
|---|---|
| author | Wang, Haomiaomiao Ward, Tomás E Zhang, Lili |
| author_facet | Wang, Haomiaomiao Ward, Tomás E Zhang, Lili |
| contents | We test whether LLMs show robust decision biases. Treating models as participants in two-arm bandits, we ran 20000 trials per condition across four decoding configurations. Under symmetric rewards, models amplified positional order into stubborn one-arm policies. Under asymmetric rewards, they exploited rigidly yet underperformed an oracle and rarely re-checked. The observed patterns were consistent across manipulations of temperature and top-p, with top-k held at the provider default, indicating that the qualitative behaviours are robust to the two decoding knobs typically available to practitioners. Crucially, moving beyond descriptive metrics to computational modelling, a hierarchical Rescorla-Wagner-softmax fit revealed the underlying strategies: low learning rates and very high inverse temperatures, which together explain both noise-to-bias amplification and rigid exploitation. These results position minimal bandits as a tractable probe of LLM decision tendencies and motivate hypotheses about how such biases could shape human-AI interaction. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_07717 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Rigidity in LLM Bandits with Implications for Human-AI Dyads Wang, Haomiaomiao Ward, Tomás E Zhang, Lili Artificial Intelligence Computer Science and Game Theory Human-Computer Interaction We test whether LLMs show robust decision biases. Treating models as participants in two-arm bandits, we ran 20000 trials per condition across four decoding configurations. Under symmetric rewards, models amplified positional order into stubborn one-arm policies. Under asymmetric rewards, they exploited rigidly yet underperformed an oracle and rarely re-checked. The observed patterns were consistent across manipulations of temperature and top-p, with top-k held at the provider default, indicating that the qualitative behaviours are robust to the two decoding knobs typically available to practitioners. Crucially, moving beyond descriptive metrics to computational modelling, a hierarchical Rescorla-Wagner-softmax fit revealed the underlying strategies: low learning rates and very high inverse temperatures, which together explain both noise-to-bias amplification and rigid exploitation. These results position minimal bandits as a tractable probe of LLM decision tendencies and motivate hypotheses about how such biases could shape human-AI interaction. |
| title | Rigidity in LLM Bandits with Implications for Human-AI Dyads |
| topic | Artificial Intelligence Computer Science and Game Theory Human-Computer Interaction |
| url | https://arxiv.org/abs/2603.07717 |