AI Paper Atlas

Research map / Continual learning / Continual fine-tuning of LLMs

Mapping Post-Training Forgetting in Language Models at Scale

Jackson Harmon, Andreas Hochlehnert, Matthias Bethge, Ameya Prabhu

ICLR 2026 · Continual learning › Continual fine-tuning of LLMs · approach: Continual pretraining of language models · 9 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
scale model size in instruction tuning (Qwen2.5 3B to 32B) (ablation)instruction tuning · knowledge benchmarkschance-adjusted forgetting: 8.5 -> 6.7Improved
scale model size in domain-continual pretraining (Coder 3B to 32B) (ablation)domain-continual pretraining · knowledge benchmarkschance-adjusted forgetting: 6.8 -> 6.4 (peak 8.0 at 14B)Mixed
use math-specialized domain-continual pretrainingQwen2.5 · knowledge benchmarkschance-adjusted forgetting: 12.9 (Math 7B) vs 7.6 (Coder 7B)Hurt
reasoning training (SFT/RL) from base vs instruction tuningbase models · knowledge benchmarksbackward transferImproved
low-data reasoning SFT from instruct modelsQwen instruct models (s1.1, LIMO) · knowledge benchmarksforgetting and backward transferNo clear effect
merge post-trained model with base checkpoint (linear/SLERP) (ablation)Qwen2.5-Coder-7B, OpenThinker3-7B, OpenThinker-7B · knowledge benchmarksforgettingMixed
measure forgetting sample-wise instead of by aggregate accuracy difference (ablation)Qwen2.5-Coder-7B · domain-continual pretrainingmeasured forgettingImproved
chat-template prompting of base-initialized coder models instead of few-shot (ablation)domain-continual pretrained models · knowledge benchmarksmeasured forgettingHurt

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