AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Using Imperfect Surrogates for Downstream Inference: Design-based Supervised Learning for Social Science Applications of Large Language Models

Naoki Egami, Musashi Hinck, Brandon M. Stewart, Hanying Wei

NeurIPS 2023 · Large language models › Language, safety and interpretability · approach: LLM unlearning and benchmarks · 24 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use DSL bias-corrected pseudo-outcome instead of surrogate-only labelsLLM surrogate (flan-ul2) · class prevalence, Toxicity (n=100)bias x100: 34.8 -> 0.0Improved
use DSL instead of surrogate-only for coverageLLM surrogate (flan-ul2) · class prevalence, Power95% CI coverage: 0.0 -> 94.8Improved
use DSL instead of gold-standard-onlygold-standard sample (n=100) · class prevalence, Misinfo.RMSE x100: 5.0 -> 4.3Improved
use DSL instead of gold-standard-onlygold-standard sample (n=100) · class prevalence, ToxicityRMSE x100: 5.0 -> 4.9No clear effect
use DSL instead of supervised learning (SL)SL with cross-fitting · class prevalence, Emotioncoverage: 91.3 -> 94.0Improved
use DSL instead of SL for RMSESL with cross-fitting · class prevalence (17 datasets)RMSE: 4.3 -> 4.3 (Misinfo.)No clear effect
use DSL instead of GSO on congressional bills logistic regression (balanced, 0-shot)GPT-3 surrogate · Congressional Bills ProjectRMSE: 14% gain in RMSE (31% in five-shot)Improved
use surrogate-only labels in downstream regression with 90% accurate surrogatesimulated logistic regression · simulation95% CI coverage: coverage only 40%Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy