Unverified50% confidenceFactExact time
Matrix-Game 4.0 employs a 30-layer causal DiT, in which self-attention reads the K/V cache and cross-attention reads the text K/V, thereby reusing historical computation during frame-by-frame generation.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe agent uses an improved version of the FSRS algorithm, combining three dynamic signals—check-in time intervals, response speed when answering questions, and the user's familiarity with words—to calculate review nodes.61% similarUnverifiedMotion gating v1 compared each frame to its neighboring frames and reused the previous frame's output; on footage with slow global gradual changes, it would skip inference indefinitely, causing the inpainted region to freeze into a rectangular block.60% similarUnverifiedReAct 每次循环产生 Thought-Action-Observation 三元组，会被逐步拼接进上下文窗口，多轮循环会快速消耗模型的 Token 配额59% similarUnverifiedThe self-attention mechanism linearly transforms the input sequence into three matrices—Query, Key, and Value—with attention weights obtained by multiplying Q by the transpose of K, normalizing via Softmax, and then multiplying by V to produce the output.58% similarUnverifiedLFC (Low Framerate Compensation) maintains the equivalent refresh rate within the VRR window by repeating each frame multiple times.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1034707API
curl https://kongchang.com/api/v1/knowledge/claims/1034707MCP
get_claim(id=1034707)