Unverified50% confidenceBenchmarkExact time
On the hardest category of questions, grounded views outperformed views that simply fed formula inputs by 17.3 points.
1
Sources
50%
Confidence
Long-term
Relevance
10/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIncluding specific evaluation dimensions in prompts (such as format, word count, key points to cover, and items to exclude) can improve model output quality by an average of 30%–50% compared to vague prompts.68% similarUnverified针对金融领域微调的嵌入模型（如FinBERT）相比通用嵌入模型在财报问答任务上准确率高出15%以上67% similarUnverifiedFor financial reports containing complex tables, the gap in downstream question-answering accuracy between simple text extraction and structure-aware parsing can reach more than 30 percentage points.65% similarUnverifiedFor inference serving, a patch that is functionally correct but causes a 30% drop in throughput is actually unacceptable, and performance should be incorporated as an evaluation dimension.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/963956API
curl https://kongchang.com/api/v1/knowledge/claims/963956MCP
get_claim(id=963956)