AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Train on Validation (ToV): Fast data selection with applications to fine-tuning

Ayush Jain, Andrea Montanari, Eren Sasoglu

ICLR 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Data augmentation and generalization bias

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
select training data with ToV scores instead of random selection or max-uncertaintyMeta-Llama-3-8B instruction tuning · Slim Orca / Alpaca GPT-4 targets, n = 8x1024test log-lossImproved
use ToV instead of LESSMeta-Llama-3-8B instruction tuning · instruction tuning Exp 1-5test log-lossMixed
apply ToV when train and validation both come from Slim OrcaMeta-Llama-3-8B instruction tuning · Exp 3 (target = pool = Slim Orca)test log-lossImproved
good data selection compared with doubling the selected set size ninstruction tuning · test log-loss vs ntest log-lossImproved
use ToV on NER versus random selectionxlm-roberta-base NER · Multinerd / Ai4p targets, n = 8x1024test log-lossImproved
use max-uncertainty selection for NERxlm-roberta-base NER · NER Exp 1-6test log-lossMixed
use LESS for NER selectionxlm-roberta-base NER · NER Exp 1-6test log-lossNo clear effect
Score-only versus Score+Random selection for instruction tuning (ablation)ToV · instruction tuning Exp 1-5test log-lossMixed
Score-only versus Score+Random selection for NER (ablation)ToV · NERtest log-lossMixed
use Method B (training resumed from validation-trained model) instead of Method A for scoring (ablation)ToV on instruction tuning · Meta-Llama-3-8Btest log-lossHurt
use Method B instead of Method A for scoring (ablation)ToV on NER · xlm-roberta-basetest log-lossNo clear effect

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