Unverified60% confidenceFactExact time
Agent Tuning is a specialized direction of supervised fine-tuning (SFT), where training samples are complete task trajectories rather than traditional question-answer pairs.
2
Sources
60%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
Related Claims
UnverifiedFoundation models for agent tasks differ markedly in training objectives from models for general conversation; the former place more emphasis on tool calling, multi-step planning, error recovery and the ability to use environmental feedback, and their training data often contains a large number of stateful interaction trajectories.77% similarUnverifiedA hybrid approach was adopted for training movements: a movement library is built in advance, and the agent is only responsible for combining movements into a training plan rather than generating movement explanations in real time.77% similarUnverifiedSpecialized training for long-horizon tasks typically requires constructing supervised data containing complete task trajectories, rather than simple single-turn question-answer pairs.74% similarUnverifiedThe planner was trained via supervised fine-tuning (SFT) on outcome-selection trajectories from 140 training questions and 28 development questions.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/489725API
curl https://kongchang.com/api/v1/knowledge/claims/489725MCP
get_claim(id=489725)