MIThinker uses an automated reverse-engineering pipeline (AugR1-MI) plus two-stage SFT+RL training to produce a plug-and-play thinker that lets an MI counseling agent reach state-of-the-art competency at roughly 1/10th the compute.
actively seeking ways to overcome the ob- stacles posed by their felony conviction and are exploring alternative paths, such as entrepreneur- ship
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
MIThinker: A Plug-and-Play Policy-Optimized Thinker For Motivational Interviewing Counseling
MIThinker uses an automated reverse-engineering pipeline (AugR1-MI) plus two-stage SFT+RL training to produce a plug-and-play thinker that lets an MI counseling agent reach state-of-the-art competency at roughly 1/10th the compute.