Table 1 Items of the large language model engagement assessment in psychological therapies (LLEAP). Each prompt concluded with the instruction: ‘Please respond in JSON format: {“rating”: [insert rating between 0-100 here]}. Do not write anything else.’

From: Development and validation of large language model rating scales for automatically transcribed psychological therapy sessions

#

Facet

Prompt text

1

Effort Betw. Sessions

Please rate how much effort the patient puts into therapeutic activities between sessions on a scale from 0 (no effort) to 100 (maximum effort).

2

Homework Completion

Please rate how consistent the patient is in completing assigned therapy homework on a scale from 0 (never completes) to 100 (always completes).

3

Motivated for Homework

Please rate how motivated the patient is to complete their therapy homework on a scale from 0 (not motivated at all) to 100 (highly motivated).

4

Accountable for Homework

Please rate how accountable the patient is for their homework and practice activities on a scale from 0 (not accountable at all) to 100 (fully accountable).

5

Participation in Exercises

Please rate how often the patient participates in therapeutic exercises or activities assigned during sessions on a scale from 0 (never participates) to 100 (always participates).

6

Engages with Techniques

Please rate how actively the patient engages with therapeutic techniques outside of sessions on a scale from 0 (not engaged at all) to 100 (fully engaged).

7

Behavioral Changes

Please rate how well the patient implements behavioral changes discussed in therapy on a scale from 0 (not at all) to 100 (very well).

8

Readiness for Sessions

Please rate how ready the patient is to engage in each therapy session on a scale from 0 (not ready at all) to 100 (completely ready).