Sherpa: Teaching LLMs to Teach Adaptively
arXiv cs.AI · 7 Oct 2026, 03:58 AEST
Sherpa is a multi-turn reinforcement learning framework designed to enhance the teaching capabilities of large language models (LLMs). It trains LLMs to adapt their instruction based on distinct learning preferences, significantly improving student performance and pedagogy scores.
Key facts
- Large language models (LLMs) have become increasingly capable problem solvers, but being able to solve a problem is not the same as being able to teach it.
- To address this, we introduce Sherpa, a multi-turn reinforcement learning framework that instantiates multiple student archetypes with LLMs conditioned on distinct learning preferences and trains a teacher model to adapt its instruction by directly maximizing their learning outcomes.
- Teacher LLMs trained with Sherpa improve instructed students' performance across all archetypes by an average of 20.5 percentage points.
- Under MathTutorBench's evaluation, Sherpa raises the overall pedagogy score from 52.5% to 79.2%, indicating better teaching responses.
- Together, Sherpa trains LLM teachers to adapt to diverse simulated students and become better aligned with human teachers, paving the road towards AI tutors teaching real students.
Read aloud
Sherpa is a multi-turn reinforcement learning framework designed to enhance the teaching capabilities of large language models L L M s. It trains L L M s to adapt their instruction based on distinct learning preferences, significantly improving student performance and pedagogy scores. Large language models L L M s have become increasingly capable problem solvers, but being able to solve a problem is not the same as being able to teach it. To address this, we introduce Sherpa, a multi-turn reinforcement learning framework that instantiates multiple student archetypes with L L M s conditioned on distinct learning preferences and trains a teacher model to adapt its instruction by directly maximizing their learning outcomes. Teacher L L M s trained with Sherpa improve instructed students' performance across all archetypes by an average of twenty point five percentage points. Under MathTutorBench's evaluation, Sherpa raises the overall pedagogy score from fifty two point five percent to seventy nine point two percent, indicating better teaching responses. Together, Sherpa trains L L M teachers to adapt to diverse simulated students and become better aligned with human teachers, paving the road towards A I tutors teaching real students.
Sources in this cluster
- Sherpa: Teaching LLMs to Teach Adaptively
arXiv cs.AI · 7 Oct 2026, 03:58 AEST