# Fine-tuned network relies on generic representation to solve unseen cognitive task

arXiv: https://arxiv.org/abs/2406.18926 · HTML: https://mlsanity.com/paper/2406.18926/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 17 references that are in the map.
- First posted on arXiv 27 Jun 2024.

## Builds on (most cited first, 8 of 8)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Generative Language Modeling for Automated Theorem Proving](https://mlsanity.com/paper/2009.03393/) (2020)
- [Polysemanticity and Capacity in Neural Networks](https://mlsanity.com/paper/2210.01892/) (2022)
- [Relating transformers to models and neural representations of the hippocampal formation](https://mlsanity.com/paper/2112.04035/) (2021)
- [Gated recurrent neural networks discover attention](https://mlsanity.com/paper/2309.01775/) (2023)
- [Transformers and Cortical Waves: Encoders for Pulling In Context Across Time](https://mlsanity.com/paper/2401.14267/) (2024)

## Abstract (arXiv)

Fine-tuning pretrained language models has shown promising results on a wide range of tasks, but when encountering a novel task, do they rely more on generic pretrained representation, or develop brand new task-specific solutions? Here, we fine-tuned GPT-2 on a context-dependent decision-making task, novel to the model but adapted from neuroscience literature. We compared its performance and internal mechanisms to a version of GPT-2 trained from scratch on the same task. Our results show that fine-tuned models depend heavily on pretrained representations, particularly in later layers, while models trained from scratch develop different, more task-specific mechanisms. These findings highlight the advantages and limitations of pretraining for task generalization and underscore the need for further investigation into the mechanisms underpinning task-specific fine-tuning in LLMs.