We construct an explainable evaluation benchmark ELawForest (A total of 4,020 items) for the aggravated assault and the picking quarrels and provoking trouble, which mainly includes four aspects: crime constitution, circumstance for sentencing, criminal pattern and sentence. The ELawForest can be used to evaluate the interpretability of the LJP outputs of the large language models hierarchically and comprehensively. We have only shown 400 annotated data so far. If you need all the data, please contact [email protected]
ashleyleeb / elawforest Goto Github PK
View Code? Open in Web Editor NEWThe evaluation benchmarks for the paper ELawForest: Explainable legal judgement prediction benchmark based on conceptual forest