Psychotherapy Dialogue Response Faithfulness Evaluation Dataset

#response faithfulness evaluation #claim-level evidence verification #dialogue consistency assessment #psychotherapy dialogue evaluation #mental health AI assessment #therapeutic response quality assurance

5000
0.3GB
JSON
CATL
MOBIUSI INC

Updated:2026-11-05

AI Analysis & Value Prop

This dataset contains psychotherapy session transcripts, model responses to evaluate, and verifiable claims extracted from those responses with annotations indicating whether each claim is supported by the client's statements, contradicts the dialogue, or goes beyond the available evidence. Its dialogue text and claim-level annotations help measure how faithfully model responses are grounded in session content. It is designed for psychotherapy-domain model evaluation, response quality review, and trustworthiness research.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
claims	array	Verifiable claims extracted from the model response, each with a groundedness label and supporting annotation details.
model_response	string	The model response whose claims are assessed against the conversation transcript.
conversation_transcript	string	The source psychotherapy dialogue between the client and therapist.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: Each record centers on a model response to a psychotherapy dialogue and includes the session transcript, the response, and verifiable claims extracted from it with annotations. The labels indicate whether claims are supported by the dialogue, contradict it, or lack evidence in the transcript.

Which capabilities can this dataset evaluate?: It can evaluate whether a model grounds its responses in session content, including whether it identifies supported claims, detects information that conflicts with the dialogue, and flags statements that go beyond the available evidence. It can also support quality reviews and trustworthiness research for psychotherapy-related models.

What are the dataset size and delivery specifications?: The available information does not specify the number of samples, file format, or data splits. The listed fields cover session transcripts, model responses, and claims with grounding annotations. Confirm the full delivery specifications with the product listing or provider before purchase or integration.

How can the dataset be obtained, and what are the license terms?: Availability and permitted use depend on the product listing and applicable license agreement. Confirm whether research, evaluation, or commercial use is allowed and review any data-use restrictions before use. This dataset supports model evaluation and research; it is not a substitute for professional psychotherapy or clinical judgment.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Psychotherapy Dialogue Response Faithfulness Evaluation Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/11bb06dfc3c8106443baeaf8b7ca7106},
  urldate={2026-10-12},
  keywords={psychotherapy dialogue dataset,psychotherapy model evaluation,dialogue faithfulness evaluation,claim evidence verification,model response evaluation,mental health AI,dialogue contradiction detection,unsupported claim detection},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task