gouki510/Wic_data_for_SAE-Eval

Dataset Card for PS-Eval Dataset Dataset Summary The PS-Eval Dataset is a suite of polysemous and monosemous contexts extracted and filtered from the WiC dataset. It aims to evaluate the ability of Sparse Autoencoders (SAEs) to disentangle polysemantic activations into monosemantic features within large language models (LLMs). The dataset contains 1,112 samples balanced between two classes: Poly-contexts: Target words with different meanings across two contexts… See the full description on the dataset page: https://huggingface.co/datasets/gouki510/Wic_data_for_SAE-Eval.