AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

Interactive Multi-fidelity Learning for Cost-effective Adaptation of Language Model with Sparse Human Supervision

Jiaxin Zhang, Zhuohang Li, Kamalika Das, Sricharan Kumar

NeurIPS 2023 · Large language models › Instruction and fine-tuning · approach: Efficient LLM fine-tuning systems · 4 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
combine 200 human annotations with 800 GPT-3.5 annotations using IMFL instead of using 600 human annotationsdolly-v2-3b · FPB, Headline, PubMedQA, and MedQAtask score: FPB: 43.16 -> 47.88; Headline: 74.13 -> 81.09; PubMedQA: 70.15 -> 73.76; MedQA: 58.31 -> 67.98Improved
use exploration-exploitation query rather than random query (ablation)IMFL · FPB, Headline, PubMedQA, and MedQAtask score: FPB: 41.94 -> 47.88; Headline: 74.32 -> 81.09; PubMedQA: 66.03 -> 73.76; MedQA: 63.77 -> 67.98Improved
retrieve in-context examples by similarity rather than randomly (ablation)IMFL with five mini-batches and variable batch size · HeadlineAverage F-1: 73.98 -> 81.09Improved
retrieve in-context examples by similarity rather than randomly (ablation)IMFL with five mini-batches and variable batch size · PubMedQAAccuracy: 67.44 -> 73.76Improved
use variable rather than equal batch sizes (ablation)IMFL with similarity-based retrieval and five mini-batches · HeadlineAverage F-1: 80.28 -> 81.09Improved
use variable rather than equal batch sizes (ablation)IMFL with similarity-based retrieval and five mini-batches · PubMedQAAccuracy: 72.05 -> 73.76Improved
use five interactive mini-batches rather than one full batch (ablation)IMFL with similarity-based retrieval · HeadlineAverage F-1: 75.48 -> 81.09Improved
use five interactive mini-batches rather than one full batch (ablation)IMFL with similarity-based retrieval · PubMedQAAccuracy: 68.90 -> 73.76Improved
increase human annotations from 200 to 400 while reducing GPT-3.5 annotations from 800 to 600 (ablation)IMFL · FPBF-1: 47.88 ± 0.98 -> 40.23 ± 1.99Hurt
reduce human annotations from 200 to 100 (ablation)IMFL · FPBF-1: 47.88 ± 0.98 -> 43.66 ± 1.42Hurt
reduce human annotations from 200 to 50 (ablation)IMFL · FPBF-1: 47.88 ± 0.98 -> 40.76 ± 1.48Hurt
use GPT-3.5 retrieval prompting rather than GPT-3.5 zero-shot annotation (ablation)GPT-3.5 annotator · MedQAannotation accuracy: 50.82 -> 59.45Improved

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