Unverified50% confidenceFactExact time
By scanning attention heads, the developer found that an induction head drives the loop, with an induction score of 0.425, while the ordinary copying score is only 0.012.
1
Sources
50%
Confidence
Long-term
Relevance
11/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedOn dependency tasks, memory raised the success rate from a leakage-verified baseline of 0.00 to 0.55–1.00.66% similarUnverifiedpr-witness improved precision from 0.67 in the first version to 0.80 and then to 0.89 through the crossrun mechanism, while achieving zero false negatives.65% similarUnverified研究者提出了「确定性注意力Transformer」(Deterministic Attention-Transformer),并在NVIDIA H100 GPU上实测到每token仅消耗0.63焦耳的能效表现64% similarUnverifiedAfter replacing the flat linear head with a label-correlation model, the detection rate improved from 0.621 to 0.682.64% similarUnverifiedAt a similar flagging rate, a zero-shot decision-model monitor caught only 26% of leaking calls, versus 82% for the linear probe.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1164436API
curl https://kongchang.com/api/v1/knowledge/claims/1164436MCP
get_claim(id=1164436)