Unverified50% confidenceFactExact time
CLA-GRPO adopts visual-type group formation, grouping claims by attribute category (color, counting, spatial relations, text, etc.).
1
Sources
50%
Confidence
Long-term
Relevance
10/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedSPLIT-RL proposes Claim-Level Advantage (CLA-GRPO), which decomposes rollouts from the visual reasoning stage into atomic visual claims and provides fine-grained advantage signals for each claim individually.70% similarUnverifiedEach layer in the layers array of a Ming Image prompt contains four key fields: description, coordinates, hierarchy_and_relation, and color_specs.63% similarUnverified视线流通常遵循F型或Z型模式，信息分级通过字号、颜色、间距等视觉权重体现内容重要性62% similarUnverifiedDesign tokens are the most atomic visual attribute units in a design system, representing visual properties in key-value pair form, such as primary-color, border-radius-sm, and font-size-body.62% similarUnverifiedEffective prompts in image generation typically include multiple dimensions such as subject description, style reference, lighting and composition, and image quality keywords.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/990752API
curl https://kongchang.com/api/v1/knowledge/claims/990752MCP
get_claim(id=990752)