Probing Chemical Language Models: Effects of Pre-training and Fine-tuning

Chemical language models (CLMs) are trained with linearized representations such as SMILES, yet it remains unclear which chemically meaningful substructures they encode. To foster a better understanding of CLMs, we conduct a systematic study and probe for 78 molecular substructures across eight pre-trained and six randomly initialized models. We furthermore study how fine-tuning on chemical downstream tasks affects the learned representations of molecular substructures. Our results show that pre-training generally improves molecular structure awareness of CLMs, particularly in the upper layers. Moreover, randomly initialized models already encode ring structures well in the first layer. Our analysis on two chemical downstream tasks further reveals that, interestingly, fine-tuning affects task-relevant molecular substructures more than others, indicating that the changes in the representations follow chemical theory.

Fork-Think with Confidence

Parallel thinking has enjoyed great success for boosting LLM performance on reasoning tasks without the need for any re-training. However, existing methods follow a think-first-then-decide paradigm, i.e., they first sample multiple reasoning paths, which inevitably leads to overgeneration, then prune or stop unnecessary paths to compensate. In contrast, decide-first-then-think, i.e., first identifying points that are likely to lead to desirable generations, has been underexplored so far. Following this paradigm, we propose Fork-think with confidence, that first identifies forking points using model confidence in a single seeding path, then triggers thinking, sampling multiple continuations and aggregating them for the final response. Our experiments across three models and three reasoning benchmarks show that Fork-think reduces the token consumption by up to 30% and run-time by up to 57%, while performing comparable to or better than parallel thinking. Our analysis reveals that Fork-think is able to identify forking points that are meaningful with respect to the downstream task and that sampling at later positions can lead to substantially better generations. Finally, we demonstrate how combining Fork-think with existing mechanisms such as early stopping and weighted voting can further boost the performance and perform comparably to existing state-of-the-art methods, without requiring any warm-up or offline training. Our results establish pre-determined forking as a promising research direction for efficient LLM reasoning.

Sparse Mixture-of-Experts Reward Models Learn Interpretable and Specialized Experts for Personalized Preference Modeling

Preference modeling plays a central role in reinforcement learning from human feedback (RLHF), enabling large language models (LLMs) to align with human values. However, most existing approaches assume a universal reward function, neglecting the diversity and heterogeneity of human preferences. To address this limitation without additional annotation costs, recent work has proposed learning multiple preference components from binary data and combining them to model individual preferences. Nevertheless, these components often fail to capture coherent and disentangled patterns, limiting their interpretability and effectiveness for personalization. In this work, we propose a sparse Mixture-of-Experts (MoE) reward model that encourages sparse routing and expert diversity during training on binary preference data. Across controlled and real-world experiments, sparse MoE learns interpretable routing patterns and specialized experts. It also improves test-time personalization, and post-adaptation shifts in expert weights provide a qualitative lens for analyzing how the model adapts to personalized preferences.

AuthorMix: Modular Authorship Style Transfer via Layer-wise Adapter Mixing

The task of authorship style transfer involves rewriting text in the style of a target author while preserving the meaning of the original text. Existing style transfer methods train a single model on large corpora to model all target styles at once: this high-cost approach offers limited flexibility for target-specific adaptation, and often sacrifices meaning preservation for style transfer. In this paper, we propose AuthorMix: a lightweight, modular, and interpretable style transfer framework. We train individual, style-specific LoRA adapters on a small set of high-resource authors, allowing the rapid training of specialized adaptation models for each new target via learned, layer-wise adapter mixing, using only a handful of target style training examples. AuthorMix outperforms existing, SoTA style-transfer baselines -- as well as GPT-5.1 -- for low-resource targets, achieving the highest overall score and substantially improving meaning preservation.

Bridging Fairness and Explainability: Can Input-Based Explanations Promote Fairness in Hate Speech Detection?

Natural language processing (NLP) models often replicate or amplify social bias from training data, raising concerns about fairness. At the same time, their black-box nature makes it difficult for users to recognize biased predictions and for developers to effectively mitigate them. While some studies suggest that input-based explanations can help detect and mitigate bias, others question their reliability in ensuring fairness. Existing research on explainability in fair NLP has been predominantly qualitative, with limited large-scale quantitative analysis. In this work, we conduct the first systematic study of the relationship between explainability and fairness in hate speech detection, focusing on both encoder- and decoder-only models. We examine three key dimensions: (1) identifying biased predictions, (2) selecting fair models, and (3) mitigating bias during model training. Our findings show that input-based explanations can effectively detect biased predictions and serve as useful supervision for reducing bias during training, but they are unreliable for selecting fair models among candidates.

B-cos LM: Efficiently Transforming Pre-trained Language Models for Improved Explainability

Post-hoc explanation methods for black-box models often struggle with faithfulness and human interpretability due to the lack of explainability in current neural models. Meanwhile, B-cos networks have been introduced to improve model explainability through architectural and computational adaptations, but their application has so far been limited to computer vision models and their associated training pipelines. In this work, we introduce B-cos LMs, i.e., B-cos networks empowered for NLP tasks. Our approach directly transforms pre-trained language models into B-cos LMs by combining B-cos conversion and task fine-tuning, improving efficiency compared to previous B-cos methods. Our automatic and human evaluation results demonstrate that B-cos LMs produce more faithful and human interpretable explanations than post hoc methods, while maintaining task performance comparable to conventional fine-tuning. Our in-depth analysis explores how B-cos LMs differ from conventionally fine-tuned models in their learning processes and explanation patterns. Finally, we provide practical guidelines for effectively building B-cos LMs based on our findings. Our code is available at https://anonymous.4open.science/r/bcos_lm.

Constrained C-Test Generation via Mixed-Integer Programming

This work proposes a novel method to generate C-Tests; a deviated form of cloze tests (a gap filling exercise) where only the last part of a word is turned into a gap. In contrast to previous works that only consider varying the gap size or gap placement to achieve locally optimal solutions, we propose a mixed-integer programming (MIP) approach. This allows us to consider gap size and placement simultaneously, achieving globally optimal solutions, and to directly integrate state-of-the-art models for gap difficulty prediction into the optimization problem. A user study with 40 participants across four C-Test generation strategies (including GPT-4) shows that our approach (MIP) significantly outperforms two of the baseline strategies (based on gap placement and GPT-4); and performs on-par with the third (based on gap size). Our analysis shows that GPT-4 still struggles to fulfill explicit constraints during generation and that MIP produces C-Tests that correlate best with the perceived difficulty. We publish our code, model, and collected data consisting of 32 English C-Tests with 20 gaps each (totaling 3,200 individual gap responses) under an open source license.

Surveying (Dis)Parities and Concerns of Compute Hungry NLP Research

Many recent improvements in NLP stem from the development and use of large pre-trained language models (PLMs) with billions of parameters. Large model sizes makes computational cost one of the main limiting factors for training and evaluating such models; and has raised severe concerns about the sustainability, reproducibility, and inclusiveness for researching PLMs. These concerns are often based on personal experiences and observations. However, there had not been any large-scale surveys that investigate them. In this work, we provide a first attempt to quantify these concerns regarding three topics, namely, environmental impact, equity, and impact on peer reviewing. By conducting a survey with 312 participants from the NLP community, we capture existing (dis)parities between different and within groups with respect to seniority, academia, and industry; and their impact on the peer reviewing process. For each topic, we provide an analysis and devise recommendations to mitigate found disparities, some of which already successfully implemented. Finally, we discuss additional concerns raised by many participants in free-text responses.

Lessons Learned from a Citizen Science Project for Natural Language Processing

Many Natural Language Processing (NLP) systems use annotated corpora for training and evaluation. However, labeled data is often costly to obtain and scaling annotation projects is difficult, which is why annotation tasks are often outsourced to paid crowdworkers. Citizen Science is an alternative to crowdsourcing that is relatively unexplored in the context of NLP. To investigate whether and how well Citizen Science can be applied in this setting, we conduct an exploratory study into engaging different groups of volunteers in Citizen Science for NLP by re-annotating parts of a pre-existing crowdsourced dataset. Our results show that this can yield high-quality annotations and attract motivated volunteers, but also requires considering factors such as scalability, participation over time, and legal and ethical issues. We summarize lessons learned in the form of guidelines and provide our code and data to aid future work on Citizen Science.

Transformers with Learnable Activation Functions

Activation functions can have a significant impact on reducing the topological complexity of input data and therefore improve the performance of the model. Selecting a suitable activation function is an essential step in neural model design. However, the choice of activation function is seldom discussed or explored in Transformer-based language models. Their activation functions are chosen beforehand and then remain fixed from pre-training to fine-tuning. As a result, the inductive biases they imposed on models cannot be adjusted during this long life cycle. Moreover, subsequently developed models (e.g., RoBERTa, BART, and GPT-3) often follow up prior work (e.g., BERT) to use the same activation function without justification. In this paper, we investigate the effectiveness of using Rational Activation Function (RAF), a learnable activation function, in the Transformer architecture. In contrast to conventional, predefined activation functions, RAFs can adaptively learn optimal activation functions during training according to input data. Our experiments show the RAF-based Transformer (RAFT) achieves a lower validation perplexity than a vanilla BERT with the GELU function. We further evaluate RAFT on downstream tasks in low- and full-data settings. Our results show that RAFT outperforms the counterpart model across the majority of tasks and settings. For instance, RAFT outperforms vanilla BERT on the GLUE benchmark by 5.71 points on average in low-data scenario (where 100 training examples are available) and by 2.05 points on SQuAD in full-data setting. Analysis of the shapes of learned RAFs further unveils that they substantially vary between different layers of the pre-trained model and mostly look very different from conventional activation functions. RAFT opens a new research direction for analyzing and interpreting pre-trained models according to the learned activation functions.