Unverified50% confidenceFactExact time
Scientific chart reasoning is a highly difficult benchmark for evaluating multimodal models, with typical tasks including reading the slope of a line chart's trend, comparing the relative heights of bars in a bar chart, parsing the cluster structure of a scatter plot, and performing multi-step numerical inference combining legends and axes.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedMathematical reasoning benchmarks include MATH and AIME, which test symbolic derivation and multi-step calculation.71% similarUnverifiedDuring inference, the model processes input triplets group by group in an autoregressive manner along the depth axis.69% similarUnverifiedThe model in this study fixes features and inequality directions, and is used to analyze the learning of decision stump thresholds68% similarUnverifiedReasoning enhancement and hypothesis generation, through chain-of-thought reasoning, tool invocation, and retrieval augmentation, enable models to construct testable scientific arguments.68% similarUnverified将多个画面组合为单张网格图输入模型，是一种利用模型空间推理能力的技巧，模型能感知网格中各格之间的位置关系66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1047314API
curl https://kongchang.com/api/v1/knowledge/claims/1047314MCP
get_claim(id=1047314)