Unverified50% confidenceFactExact time
Hide and Seek (HnS) is a concept-level attribution method that requires neither access to model weights nor a concept-annotated dataset.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedVLM-Seek incorporates a None mechanism that directly returns a not-found result when no target matching the description exists in the candidate regions, rather than forcibly drawing an incorrect bounding box.60% similarUnverified将推理链隐藏在API后端并不等于真正保护它，只要模型行为可被外部观测和大量查询就存在被逆向工程的空间60% similarUnverifiedZero retention does not equal zero learning—the fact that data is not stored does not mean that patterns and information extracted from it will not influence the model in some form55% similarUnverifiedOpenAI directly discards historical reasoning data in multi-turn conversations, neither retaining nor returning the chain of thought, and therefore presents no target that can be distilled and stolen.55% similarUnverifiedSearch grounding refers to a model retrieving external knowledge sources in real time during content generation and using the retrieved results as contextual constraints to reduce hallucinations.54% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1001892API
curl https://kongchang.com/api/v1/knowledge/claims/1001892MCP
get_claim(id=1001892)