Unverified50% confidenceFactExact time
A potential function is a scoring function in SMC that defines "how close the current partial sequence is to the goal"; it need not be differentiable or trained, and only needs to be computable in code.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe design of SMC is inspired by Speculative Decoding technology and elevates it from the token level to the action-chain level.65% similarUnverifiedSMC-based inference-time power sampling can improve LLM reasoning without any post-training.65% similarUnverifiedDuring generation, SMC maintains a set of weighted candidate sequences (particles), reweighting and resampling them at each step according to the potential functions.64% similarUnverifiedThe target score identity expresses the score function as a self-consistent relation between the current model prediction and the gradient of the unnormalized density, making it possible to construct a training objective without external training data.63% similarUnverifiedSA (Speculative Action) only performs single-step speculation, whereas SMC achieves reuse of multi-step speculative execution through the Macro Library, thereby delivering additional gains beyond single-step speculation.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1115490API
curl https://kongchang.com/api/v1/knowledge/claims/1115490MCP
get_claim(id=1115490)