Unverified50% confidenceSolutionExact time
The researchers eliminated the dependency on SenseNova at inference time by training a distilled student network with the architecture 5120→512→512→5120 (RMS normalization + SiLU activation) to reconstruct the projected teacher representations from H3's L49 conditions.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
Unverified卡内基梅隆大学开发的Cognitive Tutor引入产生式规则模型和学生模型，在代数教学中取得约0.4 SD的持续效果，被美国数千所学校部署68% similarUnverifiedThe distillation in this project treats unstructured expert human experience as the teacher, converting tacit knowledge into explicit executable rules via LLM, which differs from traditional model distillation (large model teaching a small model).66% similarUnverifiedTraining employs unrolled horizon curriculum learning, starting from K=2 and progressively doubling to 512 (2→4→8→16→32→64→128→256→512).65% similarUnverified斯坦福 CS231n 课程作业明确要求用 NumPy 从零实现神经网络65% similarUnverifiedMIT 6.390 'Introduction to Machine Learning,' formerly numbered 6.036, is taught by professors including Leslie Kaelbling and Tomás Lozano-Pérez, and its lecture notes cover complete content from linear classifiers and neural networks to reinforcement learning.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/911894API
curl https://kongchang.com/api/v1/knowledge/claims/911894MCP
get_claim(id=911894)