Popular repositories Loading

1. llm_surprisal llm_surprisal Public

   Surprisal calculation using HuggingFace LMs ("Frequency Explains the Inverse Correlation of Large Language Models’ Size, Training Data Amount, and Surprisal’s Fit to Reading Times," EACL24)

   Python 10 2

2. acl21_semproc acl21_semproc Public

   Incremental left-corner parser and pretrained model for surprisal estimation ("Surprisal Estimators for Human Reading Times Need Character Models," ACL21)

   C++ 4

3. attn_dist attn_dist Public

   Norm-aware versions of GPT-2 attention and predictors of processing difficulty calculated from them ("Entropy- and Distance-Based Predictors From GPT-2 Attention Patterns Predict Reading Times Over…

   Python 4

4. slm_surprisal slm_surprisal Public

   LM training using GPT-NeoX ("Transformer-Based Language Model Surprisal Predicts Human Reading Times Best with About Two Billion Training Tokens," Findings of EMNLP23)

   Python 2

5. cmcl21_st cmcl21_st Public

   Python 1

6. llm_decomposition llm_decomposition Public

   Linear decomposition of OPT representations and associated importance measure ΔLP ("Token-wise Decomposition of Autoregressive Language Model Hidden States for Analyzing Model Predictions," ACL23)

   Python 1