Unverified50% confidenceFactExact time
Existing long-chain reasoning methods typically rely on a single strategy and are trained with outcome-reward reinforcement learning, but long-horizon planning failures occur from time to time, and the model's implicit reasoning plan gradually drifts away from effective strategies.
1
Sources
50%
Confidence
Long-term
Relevance
11/3/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedWhen general reasoning paradigms fail in a specific domain, constraining the reasoning process with domain structure and then distilling reasoning trajectories with a teacher model may be an effective path for small models to acquire specialized capabilities.76% similarUnverifiedDecoupling strategy generation from strategy execution brings attributability, making it possible to determine when a strategy loses money whether the problem lies with the strategy itself rather than random model behavior.76% similarUnverifiedTraining models with reinforcement learning for parallel thinking has the fundamental flaw of mode collapse, namely that the multiple parallel paths output by the model degenerate into highly similar trajectories.75% similarUnverifiedPredictive relationships become invalid as platform formats change, so models need continuous re-estimation rather than one-time training for long-term use.75% similarUnverifiedA common heuristic is to train a model to predict outcomes from decisions and use its coefficients or feature importance as a proxy for the influence of each decision, but the authors show this does not always hold, and estimates become distorted when the model is misspecified or has insufficient predictive power.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1129342API
curl https://kongchang.com/api/v1/knowledge/claims/1129342MCP
get_claim(id=1129342)