Unverified50% confidenceSolutionExact time
The ASTEROID training pipeline consists of two steps: the student model collects online interaction trajectories relying only on historical observations; the privileged expert, with access to the true hidden state, provides action labels based on the student's history context for distillation.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe main technical approaches for addressing progressive drift include conversation-level anomaly detection, policy memory training based on reinforcement learning, and external stateful audit agents.69% similarUnverifiedSLVR adopts a two-stage training strategy: the first step masks text that leaks the answer and uses visual distractors to force the model to rely on the image, and the second step applies structured stage-wise supervision.69% similarVerifiedOnline policy distillation (OPD) has the student model generate its own training responses, which are then supervised by the teacher model.68% similarUnverifiedThe training data for on-policy distillation is generated by the interaction of the student's current policy with the environment, and the state distribution changes as the student policy is updated, in line with the DAgger idea, which can mitigate distribution shift.68% similarUnverifiedFoundation 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.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1084014API
curl https://kongchang.com/api/v1/knowledge/claims/1084014MCP
get_claim(id=1084014)