SELECT-LLM is the first active model selection framework for LLMs that uses expected information gain from pairwise output similarities to minimize required annotations, reporting up to 84.78% cost reduction across 23 datasets and 156 models.
Weinberger, and Yoav Artzi
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
verdicts
UNVERDICTED 4representative citing papers
Geometry-calibrated conformal abstention lets language models abstain from uncertain queries with finite-sample guarantees on both participation rate and conditional correctness of answers.
LLM digital twins of individuals achieve only modest accuracy gains over base models (weak average correlation r=0.20) and exhibit five distortions: insufficient individuation, stereotyping, representation bias, ideological bias, and hyper-rationality.
A survey deriving a unified policy gradient framework for LLM post-training methods and providing technical comparisons of PPO, GRPO, DPO variants.
citing papers explorer
-
Large Language Model Selection with Limited Annotations
SELECT-LLM is the first active model selection framework for LLMs that uses expected information gain from pairwise output similarities to minimize required annotations, reporting up to 84.78% cost reduction across 23 datasets and 156 models.
-
Geometry-Calibrated Conformal Abstention for Language Models
Geometry-calibrated conformal abstention lets language models abstain from uncertain queries with finite-sample guarantees on both participation rate and conditional correctness of answers.
-
Digital Twins as Funhouse Mirrors: Five Key Distortions
LLM digital twins of individuals achieve only modest accuracy gains over base models (weak average correlation r=0.20) and exhibit five distortions: insufficient individuation, stereotyping, representation bias, ideological bias, and hyper-rationality.
-
Reinforcement Learning for LLM Post-Training: A Survey
A survey deriving a unified policy gradient framework for LLM post-training methods and providing technical comparisons of PPO, GRPO, DPO variants.