InTraVisTo: Inside Transformer Visualisation Tool
Preprint
Abstract
The reasoning capabilities of Large Language Models (LLMs) have increased greatly over the last few years, as have their size and complexity. Nonetheless, the use of LLMs in production remains challenging due to their unpredictable nature and discrepancies that can exist between their desired behavior and their actual model output. In this paper, we introduce a new tool, InTraVisTo (Inside Transformer Visualisation Tool), designed to enable researchers to investigate and trace the computational process that generates each token in a Transformer-based LLM. InTraVisTo provides a visualization of both the internal state of the Transformer model (by decoding token embeddings at each layer of the model) and the information flow between the various components across the different layers of the model (using a Sankey diagram). With InTraVisTo, we aim to help researchers and practitioners better understand the computations being performed within the Transformer model and thus to shed some light on internal patterns and reasoning processes employed by LLMs.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Evaluating large language models as agents in the clinicnpj Digital Medicine, 2024
- LM Transparency Tool: Interactive Tool for Analyzing Transformer Language Models2024
- Survey of Hallucination in Natural Language GenerationACM Computing Surveys, 2023
- Future Lens: Anticipating Subsequent Tokens from a Single Hidden State2023
- Training Compute-Optimal Large Language Models2022
- Dodrio: Exploring Transformer Models with Interactive Visualization2021
- A Multiscale Visualization of Attention in the Transformer Model2019
- "Why Should I Trust You?"2016