4arXiv cs.CL (Computation and Language)·
Cross-source adaptation study finds simpler LLM strategies transfer more reliably for climate disclosure classification
A new arXiv preprint reframes climate disclosure classification as a cross-source adaptation problem, evaluating three LLM adaptation strategies — definitions, few-shot examples, and fine-tuning — across eleven open- and closed-source models using two corpora from different document sources (annual reports, press releases, earnings calls). The key finding is that stronger in-source strategies (similarity-based retrieval, LoRA fine-tuning) lose their advantage under source shift, while simpler approaches like random few-shot selection and well-matched definitions transfer more consistently. The result has practical implications for NLP practitioners deploying classifiers across heterogeneous document types.
Related guides (2)
Related events (8)
Cross-lingual in-context learning source language selection challenges fine-tuning assumptions
A new arXiv paper conducts a broad empirical study of cross-lingual transfer in few-shot in-context learning (ICL), spanning seven tasks, six models, and a typologically diverse set of languages. The study finds that conventional heuristics from supervised fine-tuning — such as relying on linguistic similarity or data availability — do not consistently transfer to the ICL regime. The authors also analyze language confusion as a key obstacle in generative cross-lingual ICL and propose alternative heuristics for source language selection.
Localized LoRA adaptation reveals distinct learning geometries across transformer layers and task types
A new arXiv preprint introduces a controlled benchmark spanning five learning objectives (lexical binding, factual association, behavioral policy, causal mapping, procedural reasoning) to study how adaptation site shapes what transformers learn, how well it generalizes, and how selectively it applies. The authors define each objective's 'adaptation geometry' as its profile under full-stack versus early-, middle-, or late-layer LoRA. Key findings include that factual association favors late-layer adaptation, behavioral learning separates late-layer acquisition from middle-layer gating, and causal/procedural transfer benefits most from middle or full-stack updates. Results replicate across five model families under parameter-matched controls, establishing adaptation site as a meaningful design variable.
Adversarial robustness and safety alignment in multilingual multimodal LLMs: cross-lingual vulnerability and 'safety-by-failure'
A systematic study evaluates adversarial robustness and safety alignment of multimodal LLMs across 12 languages, finding that adversarial images optimized in one language transfer to others (cross-lingual transferability). The paper introduces the concept of 'safety-by-failure': low-resource languages appear safer not due to genuine alignment but because models fail to comprehend harmful instructions in those languages. Models like Qwen3-VL that integrate multilingual capability throughout training (rather than only at instruction tuning) show genuine cross-lingual safety with active refusal. The findings challenge the assumption that low-resource language safety metrics reflect real alignment.
RL-trained LLMs learn retriever-specific query formulation strategies for RAG
A new arXiv paper presents the first systematic study of using reinforcement learning to teach LLMs to adapt query formulation strategies to different retrieval backends. The authors find that different retrievers have surprisingly distinct optimal query styles (e.g., descriptive vs. question-like), making cross-retriever strategy transfer ineffective. They introduce a branching-based rollout technique to stabilize training over multi-step retrieval trajectories and show gains from retriever-specific human guidance and model scaling.
Towards Reliable Multilingual LLMs-as-a-Judge: An Empirical Study
This paper systematically investigates strategies for extending LLM-based automatic evaluation (LLMs-as-a-Judge) to multilingual settings, covering high-, mid-, and low-resource languages (English, Spanish, Basque). The authors compare instruction translation, monolingual vs. multilingual supervision, and model size, finding that fine-tuned smaller models can match proprietary models when in-domain data is available, while zero-shot larger models are preferable out-of-domain. Two meta-evaluation datasets are extended to Spanish and Basque, and all data and code are publicly released.
Empirical study of LLM medical domain adaptation trade-offs in French QA
Researchers present a systematic comparison of continual pretraining (CPT), supervised fine-tuning (SFT), and their combination for adapting LLMs to French medical question answering. The study spans three model families, multiple sizes, and three initialization types, evaluating both multiple-choice and open-ended QA formats. Key findings: CPT+SFT yields the best MCQA scores but gains over SFT alone are often not statistically significant, making SFT a cost-effective default; for open-ended QA, CPT improves overlap metrics while SFT degrades generation quality. Cross-lingual transfer from French adaptation to English benchmarks is also demonstrated.
Inference-time steering methods compared for cross-lingual factual consistency in LLMs
Researchers investigate four intervention strategies for reducing cross-lingual factual inconsistency in LLMs, where models give different factual answers depending solely on prompt language. Testing on Gemma 3 12B Instruct, they compare persona prompting, Contrastive Activation Addition (CAA), and DPO-based adapters, finding that simple persona prompting outperforms more invasive methods in efficacy, safety, and generalization. The paper also introduces a novel culturally-rooted generalization benchmark to assess whether factual interventions transfer beyond the training distribution. Results suggest cross-lingual inconsistency is partly a selection problem addressable without weight modification.
Systematic study reveals effectiveness-fluency trade-offs in LLM conditioning methods
A new arXiv paper systematically evaluates a range of LLM conditioning methods across both concept injection and removal scenarios, finding that efficient steering methods often degrade fluency significantly. A key finding is that activation steering is substantially less effective on instruction-tuned models than on base models, a previously overlooked interaction. Simple prompting and supervised fine-tuning work for concept injection but not removal, and cheap textual metrics are found to correlate well with expensive LLM-as-judge evaluations.