Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.17079 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917418068082688 |
|---|---|
| author | Star, Michelle Aquilina, Andrew Lin, Yu-Ru |
| author_facet | Star, Michelle Aquilina, Andrew Lin, Yu-Ru |
| contents | When users seek social support from chatbots, they disclose their situation gradually, yet most evaluations of supportive LLMs rely on single-turn, fully specified prompts. We introduce a multi-turn simulation framework that closes this gap. Support-seeking narratives from five Reddit communities are decomposed into ordered fragments and revealed turn by turn to a language model. Each response is coded with the Social Support Behavior Code (SSBC), an established multi-label taxonomy that captures the composition of support, rather than a single quality score. To ask whether support choices track the model's own construal of user distress, we use linear probes on hidden representations to estimate this internal signal without altering the generation context. Across two mid-scale models (Llama-3.1-8B, OLMo-3-7B) and more than 6,200 turns, support composition shifts systematically with estimated distress: teaching declines as estimated distress rises, a finding that replicates across architectures, while increases in affective and esteem-oriented strategies (such as validation) are suggestive but model-specific and rest on noisier annotations. Community context independently shapes behavior, tracking topic and discourse norms rather than demographic categories. These trajectory-level dynamics, invisible to single-turn evaluation, motivate multi-turn auditing frameworks for socially sensitive applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_17079 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Auditing Support Strategies in LLMs through Grounded Multi-Turn Social Simulation Star, Michelle Aquilina, Andrew Lin, Yu-Ru Computation and Language When users seek social support from chatbots, they disclose their situation gradually, yet most evaluations of supportive LLMs rely on single-turn, fully specified prompts. We introduce a multi-turn simulation framework that closes this gap. Support-seeking narratives from five Reddit communities are decomposed into ordered fragments and revealed turn by turn to a language model. Each response is coded with the Social Support Behavior Code (SSBC), an established multi-label taxonomy that captures the composition of support, rather than a single quality score. To ask whether support choices track the model's own construal of user distress, we use linear probes on hidden representations to estimate this internal signal without altering the generation context. Across two mid-scale models (Llama-3.1-8B, OLMo-3-7B) and more than 6,200 turns, support composition shifts systematically with estimated distress: teaching declines as estimated distress rises, a finding that replicates across architectures, while increases in affective and esteem-oriented strategies (such as validation) are suggestive but model-specific and rest on noisier annotations. Community context independently shapes behavior, tracking topic and discourse norms rather than demographic categories. These trajectory-level dynamics, invisible to single-turn evaluation, motivate multi-turn auditing frameworks for socially sensitive applications. |
| title | Auditing Support Strategies in LLMs through Grounded Multi-Turn Social Simulation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2604.17079 |