AstroAI — Center for Astrophysics | Harvard & Smithsonian
AstroAI
Developing Artificial Intelligence to Solve the Mysteries of the Universe
Explore
  • Home
  • Research
  • EarthAI
  • People
  • Events
  • Latest News
  • Lunch Talks
  • Summer Program
  • Workshop
  • Apply
  • Contact
Home AstroAI Workshop 2025
AstroAI Workshop 2025
Cancel

2025 Workshop

AstroAI Workshop 2025

Details Invited Speakers Abstracts Schedule Venue Code of Conduct
Poster

2025 workshop program

Building Honest Agents Through Introspection: Probe-driven Generation of Confidence Scores

Presented by
Rocco Di Tella
Program time
Monday, July 7th, 3:30 - 5:00 PM
01

The work

Abstract

Large language models (LLMs) are the engine behind agentic AI, providing language processing, planning, and reasoning capabilities. Unfortunately, current LLMs do not directly provide a measure of confidence for the responses they produce. This poses a serious problem for high-risk applications, where only responses that are very likely to be correct should be accepted. We propose to explore supervised approaches for computing confidence measures for answers provided by LLMs. To this end, we will develop models that probe the LLM’s internal representations to predict whether an answer is correct or not, focusing on structured architectures with strong inductive biases to facilitate generalization to unseen tasks. We will train and evaluate our models on a variety of NLP datasets, using proper scoring rules to assess performance of the produced scores.

Keep exploring

More from the 2025 workshop

Browse the complete collection of presentations or return to the full workshop program.

Browse abstracts View schedule

© 2026 AstroAI. Some rights reserved.

Powered by Jekyll with Chirpy theme.

A new version of content is available.