TH1.R1.2

Predicting Uncertainty of Generative LLMs with MARS: Meaning-Aware Response Scoring

Yavuz Faruk Bakman, Duygu Nur Yaldiz, Baturalp Buyukates, University of Southern California, United States; Chenyang Tao, Dimitrios Dimitriadis, Amazon, United States; Salman Avestimehr, University of Southern California, United States
Session:
Language Models
Track:
8: Machine Learning
Location:
Ballroom II & III
Presentation Time:
Thu, 11 Jul, 10:05 - 10:25
Session Chair:
Homa Esfahanizadeh,
Presentation
Not logged in.
Discussion
Not logged in.
Session TH1.R1
TH1.R1.1: Asymptotics of Language Model Alignment
Joy Yang, University of Sydney, Australia; Salman Salamatian, MIT, United States; Ziteng Sun, Ananda Theertha Suresh, Ahmad Beirami, Google, Australia
TH1.R1.2: Predicting Uncertainty of Generative LLMs with MARS: Meaning-Aware Response Scoring
Yavuz Faruk Bakman, Duygu Nur Yaldiz, Baturalp Buyukates, University of Southern California, United States; Chenyang Tao, Dimitrios Dimitriadis, Amazon, United States; Salman Avestimehr, University of Southern California, United States
TH1.R1.3: TexShape: Information Theoretic Sentence Embedding for Language Models
Kaan Kale, Bogazici University, Turkey; Homa Esfahanizadeh, Nokia Bell Labs, United States; Noel Elias, Oguzhan Baser, University of Texas at Austin, United States; Muriel Medard, MIT, United States; Sriram Vishwanath, University of Texas at Austin, United States
TH1.R1.4: The Likelihood Gain of a Language Model as a Metric for Text Summarization
Dana Levin, Alon Kipnis, Reichman University, Israel
Resources
No resources available.