Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks

Samyak Jain, Robert Kirk, E. Lubana, Robert P. Dick, Hidenori Tanaka +3 more
2/12/2026

No official code found. Best implementation: RobertKirk/tinystories-wrappers (9 stars, MIT, last commit 2023-12-18).

Semantic Scholar

Code Implementations(1)

RobertKirk/tinystories-wrappersCommunity58%
MIT

Code for the TinyStories experiments from "Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks".

91Shell, PythonDec 8, 20232 years agoMIT
interpretabilitylanguage-modelllama2

Similar Papers

Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks

76% match1 repo92023

Efficiency in Focus: LayerNorm as a Catalyst for Fine-tuning Medical Visual Language Pre-trained Models

76% match2024

How to prepare your task head for finetuning

75% match2023

Fine-Tuning is Subgraph Search: A New Lens on Learning Dynamics

75% match2025

Learning Dynamics of LLM Finetuning

74% match1 repo2252026

Fully Fine-tuned CLIP Models are Efficient Few-Shot Learners

74% match2024

Cite this paper

@article{jain2026mechanistically,
  title  = {Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks},
  author = {Samyak Jain and Robert Kirk and E. Lubana and Robert P. Dick and Hidenori Tanaka and Tim Rocktäschel and Edward Grefenstette and D. Krueger},
  year   = {2026},
  url    = {https://api.semanticscholar.org/CorpusID:320da764bd416e7c0dcc1b89fd16cf73560ea852},
  journal = {ICLR 2024 2024}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/320da764bd416e7c0dcc1b89fd16cf73560ea852.svg)](https://codeofpaper.com/paper/320da764bd416e7c0dcc1b89fd16cf73560ea852)

Discussion