# Learning to Act with Task Progress: Distilling Small Agents from Compact Teacher Supervision

arXiv: https://arxiv.org/abs/2610.10332 · HTML: https://mlsanity.com/paper/2610.10332/

## Key facts

- No citing papers in the map yet (data through 7 Oct 2026).
- Builds on 14 of its 20 references that are in the map.
- First posted on arXiv 7 Oct 2026.

## Builds on (most cited first, 14 of 14)

- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Group-in-Group Policy Optimization for LLM Agent Training](https://mlsanity.com/paper/2505.10978/) (2025)
- [FireAct: Toward Language Agent Fine-tuning](https://mlsanity.com/paper/2310.05915/) (2023)
- [RLVMR: Reinforcement Learning with Verifiable Meta-Reasoning Rewards for Robust Long-Horizon Agents](https://mlsanity.com/paper/2507.22844/) (2025)
- [ReAct Meets ActRe: When Language Agents Enjoy Training Data Autonomy](https://mlsanity.com/paper/2403.14589/) (2024)
- [HiPER: Hierarchical Reinforcement Learning with Explicit Credit Assignment for Large Language Model Agents](https://mlsanity.com/paper/2602.16165/) (2026)
- [Milestone-Guided Policy Learning for Long-Horizon Language Agents](https://mlsanity.com/paper/2605.06078/) (2026)
- [Structured Agent Distillation for Large Language Model](https://mlsanity.com/paper/2505.13820/) (2025)
- [SAGE-OPD: Selective Agent-Guided Intervention for Multi-Turn On-Policy Distillation](https://mlsanity.com/paper/2606.19659/) (2026)
- [From History to State: Constant-Context Skill Learning for LLM Agents](https://mlsanity.com/paper/2605.05413/) (2026)
- [Solving the Granularity Mismatch: Hierarchical Preference Learning for Long-Horizon LLM Agents](https://mlsanity.com/paper/2510.03253/) (2025)
- [EnvRL: Learn from Environment Dynamics in Agentic Reinforcement Learning](https://mlsanity.com/paper/2606.17680/) (2026)
- [Spinning Straw into Gold: Relabeling LLM Agent Trajectories in Hindsight for Successful Demonstrations](https://mlsanity.com/paper/2607.04235/) (2026)
- [Automata from Agent Traces: Failure and Next-Step Prediction](https://mlsanity.com/paper/2608.23670/) (2026)

## Abstract (arXiv)

Learning from large-model demonstrations offers a way to train small agents that can complete recurring tasks without calling a large model at every step. A central design choice is what to retain from teacher trajectories that contain reasoning, actions, and information about task progress. We introduce Task-Progress Distillation (TPD), an offline approach that pairs each demonstrated action with a short label describing the current task stage. The student learns these compact targets and selects actions by jointly scoring admissible stage--action pairs, which a deterministic harness executes in the environment. On ALFWorld, a 1.7B student trained with 404 demonstrations achieves 72.4% mean unseen task success with either TPD or action-only supervision, compared with 48.3% for a reasoning-trained student using constrained action selection. Explicit stages provide an additional benefit at 200 demonstrations, improving success from 48.0% to 67.7% over action-only supervision. With more demonstrations, the action-only student closes the gap, and both approaches reach 76.9% at 808 demonstrations. Shared-history analyses link part of TPD's local advantage to better decisions when moving between subgoals, particularly from object acquisition to processing. These results show that compact supervision can train effective small task agents, while explicit task progress provides additional guidance at an intermediate demonstration budget.