Safety
Move by Move: Measuring and Steering How LLMs Conduct Psychotherapy
arXiv:2608.21325v1 Announce Type: new Abstract: Users increasingly turn to large language models for emotional support, yet little is known about how these models actually conduct a psychotherapy inte
arXiv:2608.21325v1 Announce Type: new Abstract: Users increasingly turn to large language models for emotional support, yet little is known about how these models actually conduct a psychotherapy interaction. We introduce an ontology of ten therapeutic moves: compact, function-based categories grounded in the MULTI-60 inventory, validated through an annotation campaign with five licensed psychologists, and scaled with a judge-based approach that matches expert agreement. Applying it to real counseling transcripts and model-led sessions, we compare the move distributions between human clinicians and a panel of frontier models. Models over-use inquiry at up to three times the human rate, neglect psychoeducation, and are strongly context-anchored: they carry forward strategies initiated by a human clinician but rarely initiate them themselves. Exposing the ontology as a set of tools roughly halves the mean deviation from the human move distribution and improves turn-level alignment with human therapist by 7-9 percentage points, without any fine-tuning.
Related
- DRIFT: Decoupled Rollouts and Importance-Weighted Fine-Tuning for Efficient Multi-Turn Optimization
- Toward Culturally Aligned LLMs through Ontology-Guided Multi-Agent Reasoning
- Every Token Counts: Exact Likert-Scale Distributions for Measuring LLM Attitudes and Biases
Source: arXiv cs.CL | 2026-08-24