# Self-Specialized Teachers for Domain Post-Training

arXiv: https://arxiv.org/abs/2608.28647 · HTML: https://mlsanity.com/paper/2608.28647/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 30 of its 37 references that are in the map.
- First posted on arXiv 16 Aug 2026.

## Builds on (most cited first, 30 of 30)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [Learning without Forgetting](https://mlsanity.com/paper/1606.09282/) (2016)
- [iCaRL: Incremental Classifier and Representation Learning](https://mlsanity.com/paper/1611.07725/) (2016)
- [BioBERT: a pre-trained biomedical language representation model for biomedical text mining](https://mlsanity.com/paper/1901.08746/) (2019)
- [Gradient Episodic Memory for Continual Learning](https://mlsanity.com/paper/1706.08840/) (2017)
- [FitNets: Hints for Thin Deep Nets](https://mlsanity.com/paper/1412.6550/) (2014)
- [Don't Stop Pretraining: Adapt Language Models to Domains and Tasks](https://mlsanity.com/paper/2004.10964/) (2020)
- [SciBERT: A Pretrained Language Model for Scientific Text](https://mlsanity.com/paper/1903.10676/) (2019)
- [Memory Aware Synapses: Learning what (not) to forget](https://mlsanity.com/paper/1711.09601/) (2017)
- [Efficient Lifelong Learning with A-GEM](https://mlsanity.com/paper/1812.00420/) (2018)
- [Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks](https://mlsanity.com/paper/1506.03099/) (2015)
- [Sequence-Level Knowledge Distillation](https://mlsanity.com/paper/1606.07947/) (2016)
- [Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes](https://mlsanity.com/paper/2305.02301/) (2023)
- [An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuning](https://mlsanity.com/paper/2308.08747/) (2023)
- [Deep Mutual Learning](https://mlsanity.com/paper/1706.00384/) (2017)
- [Sequence Level Training with Recurrent Neural Networks](https://mlsanity.com/paper/1511.06732/) (2015)
- [Improved Knowledge Distillation via Teacher Assistant](https://mlsanity.com/paper/1902.03393/) (2019)
- [FinQA: A Dataset of Numerical Reasoning over Financial Data](https://mlsanity.com/paper/2109.00122/) (2021)
- [Born Again Neural Networks](https://mlsanity.com/paper/1805.04770/) (2018)
- [Do Deep Nets Really Need to be Deep?](https://mlsanity.com/paper/1312.6184/) (2013)
- [Scaling Laws for Transfer](https://mlsanity.com/paper/2102.01293/) (2021)
- [Professor Forcing: A New Algorithm for Training Recurrent Networks](https://mlsanity.com/paper/1610.09038/) (2016)
- [LAMOL: LAnguage MOdeling for Lifelong Language Learning](https://mlsanity.com/paper/1909.03329/) (2019)
- [Fine-tuned Language Models are Continual Learners](https://mlsanity.com/paper/2205.12393/) (2022)
- [More Than Catastrophic Forgetting: Integrating General Capabilities For Domain-Specific LLMs](https://mlsanity.com/paper/2405.17830/) (2024)

## Built on by (most cited first, 1 of 1)

- [From Rollouts to Recipes: Self-Contained Post-Training for LLMs](https://mlsanity.com/paper/2609.01422/) (2026)

## Abstract (arXiv)

Target-only post-training can improve performance in a specialized domain while degrading behaviors that a general-purpose base model acquired before adaptation. We study this problem when target-domain data are available but a representative replay corpus is not. We propose self-specialized teacher distillation (SSTD), a two-stage procedure that first trains a copy of the base model into a domain teacher, then distills its token distribution to a student on prefixes sampled from the student itself. Teacher training combines standard target supervision with base-aware key-token weighting and distribution alignment to the frozen base model; on-policy distillation then places domain feedback on states the student can encounter at inference time. On financial numerical reasoning, medical question answering, and legal holding identification, SSTD retains much of the target improvement of direct fine-tuning while improving the mean score on the evaluated general suite by 4.8--5.0 points at the reported operating point. The pattern persists across Qwen3 sizes and on Gemma backbones. SSTD requires neither an external teacher nor general replay data.