The paper introduces a method that adapts Contextual Decomposition for Transformers (CD‑T) to discover neural circuits without relying on clean counterfactuals, using label‑balanced activation means and task‑directional relevance scoring. These circuits are then employed in Circuit‑Targeted Supervised Fine‑Tuning (CT‑SFT), which restricts parameter updates to task‑relevant heads and LayerNorm, leading to competitive performance on low‑resource language adaptation tasks such as NusaX cross‑lingual sentiment transfer. CT‑SFT consistently avoids catastrophic forgetting and preserves source‑language and related‑task performance, offering a more controlled alternative to global fine‑tuning, as further validated on the XNLI benchmark.
arXiv:2604. 26170v2 Announce Type: replace Abstract: Adapting large language models (LLMs) to a targeted task efficiently and effectively remains a fundamental challenge.
arXiv:2607. 04733v1 Announce Type: cross Abstract: Supervised fine-tuning (SFT) is the standard approach for adapting pretrained language models to downstream domains, yet it often improves target-domain behavior at the cost of degrading pre-existing capabilities.
Supervised fine-tuning (SFT) is the standard approach for adapting pretrained language models to downstream domains, yet it often improves target-domain behavior at the cost of degrading pre-existing capabilities. Standard cross-entropy fine-tuning promotes only the observed label token and leaves unconstrained how probability mass is redistributed over other plausible alternatives, potentially distorting the rich local preference structure learned during pretraining.
arXiv:2512. 20661v2 Announce Type: replace Abstract: Transformer-based pre-trained language models (PLMs) excel in text classification but suffer from attention dilution and attention sink effects, forcing models to over-focus on task-irrelevant tokens.
CroCo introduces cross‑lingual contrastive preference tuning on self‑generations, extending prior English‑only methods to 14 high‑ and low‑resource languages. A reward model trained solely on English preferences, applied to a multilingual base, yields effective within‑language rankings and improves performance in both monolingual and multilingual settings without catastrophic forgetting. The approach requires on‑policy data; off‑policy responses and online preference optimization offer limited gains, yet on structured tasks CroCo matches or surpasses the base model in most languages, and on open‑ended generation it wins 28/30 judge evaluations across 15 languages.
arXiv:2608.29613v1 Announce Type: cross Abstract: Function vectors (FVs) have recently emerged as a promising mechanism for steering the behavior of large language models (LLMs) by injecting task-spe...
arXiv:2506. 10292v2 Announce Type: replace-cross Abstract: Training deep learning networks with minimal supervision has gained significant research attention due to its potential to reduce reliance on extensive labelled data.
arXiv:2606. 01252v1 Announce Type: cross Abstract: Multi-target cross-lingual text summarization (MTXLS), which summarizes a source document into multiple target languages, is increasingly important as users consume content in diverse languages, but remains underexplored.
arXiv:2502. 08363v3 Announce Type: replace-cross Abstract: We present Top-Theta (Top-$\theta$) Attention, a training-free method for sparsifying transformer attention during inference.
arXiv:2606. 18033v1 Announce Type: cross Abstract: Cross-lingual transfer in multilingual NLP has been widely explored in supervised fine-tuning contexts, where factors like data availability and linguistic similarity largely determine transfer quality.
arXiv:2607. 19101v1 Announce Type: cross Abstract: Reliable Text Difficulty Assessment is a prerequisite for valid text simplification workflows and personalized learning applications.
arXiv:2607. 19243v1 Announce Type: cross Abstract: Although Large Language Models (LLMs) demonstrate remarkable multilingual fluency, their internal knowledge representations remain disproportionately biased toward high-resource languages.