Unverified50% confidenceFactExact time
Self-supervised TTT work can be traced back to around 2020, enabling models to adapt to the current input by constructing auxiliary self-supervised tasks (such as rotation prediction) on test samples.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified测试时训练（TTT）在推理阶段利用当前输入的自监督信号对部分参数进行短暂局部梯度更新再执行预测78% similarVerifiedThe Needle in a Haystack test was originally proposed by Greg Kamradt in late 2023 to evaluate a model's ability to precisely retrieve specific information in long contexts.63% similarUnverifiedThe researchers tested PlanFence in 30 controlled real-time workflows that included post-plan-generation revisions.62% similarUnverifiedDetectGPT was proposed by Mitchell et al. in 2023 and determines the source of text by applying multiple random perturbations to the original text and comparing the log-probability changes before and after perturbation.61% similarUnverifiedThe Eval-Driven Development methodology of defining evaluation criteria and test sets before optimizing prompts is in the same spirit as Test-Driven Development (TDD).61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/907373API
curl https://kongchang.com/api/v1/knowledge/claims/907373MCP
get_claim(id=907373)