# Beyond Transfer Accuracy: Mechanism-Guided Controlled Adaptation for Low-Resource Languages

arXiv: https://arxiv.org/abs/2601.08146 · HTML: https://mlsanity.com/paper/2601.08146/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Circuits and mechanistic analysis, one of 13 directions in Mechanistic interpretability.
- Builds on 31 of its 43 references that are in the map.
- First posted on arXiv 13 Jan 2026.

## Placed in

- Mechanistic interpretability / [Circuits and mechanistic analysis](https://mlsanity.com/mechinterp/circuits-and-mechanistic-analysis/)

## Builds on (most cited first, 30 of 31)

- [The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks](https://mlsanity.com/paper/1803.03635/) (2018)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [Beyond Accuracy: Behavioral Testing of NLP models with CheckList](https://mlsanity.com/paper/2005.04118/) (2020)
- [XNLI: Evaluating Cross-lingual Sentence Representations](https://mlsanity.com/paper/1809.05053/) (2018)
- [The State and Fate of Linguistic Diversity and Inclusion in the NLP World](https://mlsanity.com/paper/2004.09095/) (2020)
- [XTREME: A Massively Multilingual Multi-task Benchmark for Evaluating Cross-lingual Generalization](https://mlsanity.com/paper/2003.11080/) (2020)
- [Towards Automated Circuit Discovery for Mechanistic Interpretability](https://mlsanity.com/paper/2304.14997/) (2023)
- [Mechanistic Interpretability for AI Safety -- A Review](https://mlsanity.com/paper/2404.14082/) (2024)
- [Fine-Tuning Pretrained Language Models: Weight Initializations, Data Orders, and Early Stopping](https://mlsanity.com/paper/2002.06305/) (2020)
- [On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines](https://mlsanity.com/paper/2006.04884/) (2020)
- [Polyjuice: Generating Counterfactuals for Explaining, Evaluating, and Improving Models](https://mlsanity.com/paper/2101.00288/) (2021)
- [Attribution Patching Outperforms Automated Circuit Discovery](https://mlsanity.com/paper/2310.10348/) (2023)
- [SIB-200: A Simple, Inclusive, and Big Evaluation Dataset for Topic Classification in 200+ Languages and Dialects](https://mlsanity.com/paper/2309.07445/) (2023)
- [Have Faith in Faithfulness: Going Beyond Circuit Overlap When Finding Model Mechanisms](https://mlsanity.com/paper/2403.17806/) (2024)
- [Composable Sparse Fine-Tuning for Cross-Lingual Transfer](https://mlsanity.com/paper/2110.07560/) (2021)
- [IndoNLU: Benchmark and Resources for Evaluating Indonesian Natural Language Understanding](https://mlsanity.com/paper/2009.05387/) (2020)
- [Circuit Component Reuse Across Tasks in Transformer Language Models](https://mlsanity.com/paper/2310.08744/) (2023)
- [NusaX: Multilingual Parallel Sentiment Dataset for 10 Indonesian Local Languages](https://mlsanity.com/paper/2205.15960/) (2022)
- [One Country, 700+ Languages: NLP Challenges for Underrepresented Languages and Dialects in Indonesia](https://mlsanity.com/paper/2203.13357/) (2022)
- [Overcoming Catastrophic Forgetting in Zero-Shot Cross-Lingual Generation](https://mlsanity.com/paper/2205.12647/) (2022)
- [Explorations of Self-Repair in Language Models](https://mlsanity.com/paper/2402.15390/) (2024)
- [NusaCrowd: Open Source Initiative for Indonesian NLP Resources](https://mlsanity.com/paper/2212.09648/) (2022)
- [The Same But Different: Structural Similarities and Differences in Multilingual Language Modeling](https://mlsanity.com/paper/2410.09223/) (2024)
- [Sparse Transfer Learning via Winning Lottery Tickets](https://mlsanity.com/paper/1905.07785/) (2019)
- [Efficient Automated Circuit Discovery in Transformers using Contextual Decomposition](https://mlsanity.com/paper/2407.00886/) (2024)
- [EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification](https://mlsanity.com/paper/2502.06852/) (2025)
- [Towards Understanding Fine-Tuning Mechanisms of LLMs via Circuit Analysis](https://mlsanity.com/paper/2502.11812/) (2025)
- [Rethinking Circuit Completeness in Language Models: AND, OR, and ADDER Gates](https://mlsanity.com/paper/2505.10039/) (2025)
- [LoraxBench: A Multitask, Multilingual Benchmark Suite for 20 Indonesian Languages](https://mlsanity.com/paper/2508.12459/) (2025)
- [NusaAksara: A Multimodal and Multilingual Benchmark for Preserving Indonesian Indigenous Scripts](https://mlsanity.com/paper/2502.18148/) (2025)

## Abstract (arXiv)

Existing circuit discovery methods rely on templated tasks with clean counterfactuals, limiting their use on diverse natural text. We adapt Contextual Decomposition for Transformers (CD-T) for unstructured settings via label-balanced activation means and task-directional relevance scoring, enabling counterfactual-free circuit discovery. We leverage the discovered circuits for Circuit-Targeted Supervised Fine-Tuning (CT-SFT), restricting parameter updates to task-relevant heads and LayerNorm. Experiments on NusaX cross-lingual sentiment transfer show that CT-SFT is highly competitive for low-resource adaptation. While non-circuit sparse updates and full fine-tuning sometimes match target accuracy through capacity recruitment, CT-SFT most consistently avoids catastrophic forgetting, preserving source-language and related-task performance. Extensions to XNLI support the source-retention and intervention findings on a harder task and two model families, showing that circuit-targeted adaptation provides a more controlled, intervention-supported alternative to global fine-tuning.