LLaMA-Annotate—Visualizing Token-Level Confidences for LLMs

Erik Schultheis*, S. T. John

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

Abstract

LLaMA-Annotate is a tool that allows visually inspecting the confidences that a large language model assigns to individual tokens, and the alternative tokens considered for that position. We provide both a simple, non-interactive command-line interface, as well as a more elaborate web application. Besides generally helping to form an intuition about the “thinking” of the LLM, our tool can be used for context-aware spellchecking, or to see how a different prompt or a differently trained LLM can impact the interpretation of a piece of text. The tool can be tried online at https://huggingface.co/spaces/s-t-j/llama-annotate.

Original languageEnglish
Title of host publicationMachine Learning and Knowledge Discovery in Databases. Research Track and Demo Track - European Conference, ECML PKDD 2024, Proceedings
EditorsAlbert Bifet, Povilas Daniušis, Jesse Davis, Tomas Krilavičius, Meelis Kull, Eirini Ntoutsi, Kai Puolamäki, Indrė Žliobaitė
PublisherSpringer
Pages424-428
Number of pages5
ISBN (Electronic)978-3-031-70371-3
ISBN (Print)978-3-031-70370-6
DOIs
Publication statusPublished - 1 Sept 2024
MoE publication typeA4 Conference publication
EventEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases - Vilnius, Lithuania
Duration: 9 Sept 202413 Sept 2024

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer
Volume14948 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Abbreviated titleECML PKDD
Country/TerritoryLithuania
CityVilnius
Period09/09/202413/09/2024

Keywords

  • Large language model
  • Token-level confidence
  • Visualization

Fingerprint

Dive into the research topics of 'LLaMA-Annotate—Visualizing Token-Level Confidences for LLMs'. Together they form a unique fingerprint.

Cite this