Visualizing Attention in Transformer-Based Language Representation Models
read the original abstract
We present an open-source tool for visualizing multi-head self-attention in Transformer-based language representation models. The tool extends earlier work by visualizing attention at three levels of granularity: the attention-head level, the model level, and the neuron level. We describe how each of these views can help to interpret the model, and we demonstrate the tool on the BERT model and the OpenAI GPT-2 model. We also present three use cases for analyzing GPT-2: detecting model bias, identifying recurring patterns, and linking neurons to model behavior.
This paper has not been read by Pith yet.
Forward citations
Cited by 3 Pith papers
-
Massive Activations in Large Language Models
Massive activations are constant large values in LLMs that function as indispensable bias terms and concentrate attention probabilities on specific tokens.
-
SSA: Improving Performance With a Better Scoring Function
Replacing Softmax with Scaled Signed Averaging in transformer attention improves generalization under distribution shifts for in-context learning and boosts results on NLP benchmarks.
-
Patent Claim Generation by Fine-Tuning OpenAI GPT-2
Fine-tunes GPT-2 on patent claims, probes training steps, analyzes conditional and unconditional sampling outputs, proposes a new sampling method, and releases an email bot for exploration.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.