SKR is a local, unsupervised method that improves LLM performance on financial document tasks by over 40% in retrieval recall, 76% in detection latency reduction, and 33% in anomaly AUPRC, outperforming leading models by at least 12.6%.
We report results us- ing k∈ {1,3,5,10} for our financial document datasets but using k∈ {10,15,20} for MMDocRAG to follow its metrics
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Self Knowledge Re-expression: A Fully Local Method for Adapting LLMs to Tasks Using Intrinsic Knowledge
SKR is a local, unsupervised method that improves LLM performance on financial document tasks by over 40% in retrieval recall, 76% in detection latency reduction, and 33% in anomaly AUPRC, outperforming leading models by at least 12.6%.