태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
출처에 같은 태스크 이름으로 연결된 데이터셋입니다. 동명 태스크가 포함될 수 있으므로, 비교할 때는 평가 결과의 조건도 확인하세요.
This is the dataset for knowledge editing. It contains six tasks: ZsRE, $Wiki_{recent}$, $Wiki_{counterfact}$, WikiBio, ConvSent and Sanitation. This repo shows the former 4 tasks and you can get the data for ConvSent and Sanitation from their original papers.
HalluEditBench is a comprehensive benchmark for evaluating knowledge editing methods' effectiveness in correcting real-world hallucinations. HalluEdit features a rigorously constructed dataset spanning nine domains and 26 topics. It evaluates methods across five dimensions: Efficacy, Generalization, Portability, Locality, and Robustness.
WikiFactDiff is a dataset designed as a resource to perform atomic factual knowledge updates on language models, with the goal of aligning them with current knowledge. It describes the evolution of factual knowledge between two dates, named T_old and T_new, in the form of semantic triples. To enable the possibility of evaluating knowledge algorithms (such as ROME, MEND, MEMIT, etc.), these triples are verbalized and neighbor facts are determined to check for eventual bleedover.
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.