ICL with LLMs reduces absolute imputation error for survey data versus MICE PMM across MCAR/MAR/MNAR mechanisms and yields narrower intervals with near-nominal coverage.
Demystifying prediction powered inference.arXiv preprint arXiv:2601.20819
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
citation-role summary
method 1
citation-polarity summary
years
2026 7verdicts
UNVERDICTED 7roles
method 1polarities
use method 1representative citing papers
Multi-task PPI framework uses cross-task recalibration to improve inference power across related tasks, with a proof that gains require nonlinear proxy-ground-truth structure, shown on synthetic data and a 2024 election LM audit case study.
Post-hoc calibration of miscalibrated black-box predictions on a labeled sample improves efficiency of prediction-powered inference for semisupervised mean estimation.
Proposes task exchangeability as a condition for valid inference when using synthetic data in scientific research, with methods and extensions demonstrated on surveys and AI evaluations.
Presents a unified statistical framework using the learn-then-test paradigm for hyperparameter selection that provides explicit finite-sample guarantees on application-specific reliability requirements via hypothesis testing.
Including LLM predictions as covariates in standard regression adjustment for randomized experiments reduces variance with a do-no-harm property that reverts to the unadjusted estimator when predictions are uninformative.
GLIDE is a Python library that packages multiple PPI estimators and samplers for reliable GenAI evaluation and reports annotation savings in an agentic case study.
citing papers explorer
-
In-Context Learning for the Imputation of Public Opinion Data with Large Language Models
ICL with LLMs reduces absolute imputation error for survey data versus MICE PMM across MCAR/MAR/MNAR mechanisms and yields narrower intervals with near-nominal coverage.
-
Prediction-Powered Inference Across Many Tasks for AI Evaluation & Social Science Research
Multi-task PPI framework uses cross-task recalibration to improve inference power across related tasks, with a proof that gains require nonlinear proxy-ground-truth structure, shown on synthetic data and a 2024 election LM audit case study.
-
Calibeating Prediction-Powered Inference
Post-hoc calibration of miscalibrated black-box predictions on a labeled sample improves efficiency of prediction-powered inference for semisupervised mean estimation.
-
Valid Inference with Synthetic Data via Task Exchangeability
Proposes task exchangeability as a condition for valid inference when using synthetic data in scientific research, with methods and extensions demonstrated on surveys and AI evaluations.
-
Statistically Valid Hyperparameter Selection: From Tuning to Guarantees
Presents a unified statistical framework using the learn-then-test paradigm for hyperparameter selection that provides explicit finite-sample guarantees on application-specific reliability requirements via hypothesis testing.
-
AI-Assisted Variance Reduction in Randomized Experiments
Including LLM predictions as covariates in standard regression adjustment for randomized experiments reduces variance with a do-no-harm property that reverts to the unadjusted estimator when predictions are uninformative.
-
Industrializing Prediction-Powered Inference: The GLIDE Library for Reliable GenAI and Agentic Systems Evaluation
GLIDE is a Python library that packages multiple PPI estimators and samplers for reliable GenAI evaluation and reports annotation savings in an agentic case study.