All papers

The paper introduces a CPU front-end tokenizer that encodes only session deltas to speed up multi-turn inference.

The paper studies the durability of inference-time value steering in multi-turn conversations when user and model values conflict.

The paper presents a diagnostic evaluation framework to measure how inference-time techniques affect parametric knowledge extraction in models.

The paper introduces a two-part benchmark to evaluate language models on general and examination knowledge concerning Taiwan.

The paper presents an audio evaluation benchmark assessing accuracy drops between speech-input and text-input pipelines in Southeast Asian languages.

The paper analyzes question-level variations and prompt-execution mismatches when transferring reasoning prompts across different language models.

The paper proposes a diagnostic test hiding target identity to evaluate structure-property chemical reasoning in language models.

The paper introduces a method that weights rubric criteria through pairwise model comparisons to improve evaluation accuracy.

The paper shows pruned reasoning models fail by looping, proposing an activation steering method to restore proper termination.

The paper tests a heuristic to set the number of transformer attention heads using the intrinsic dimension of token representations.

The paper systematizes grading requirements for open-text legal exams and compares eight evaluation paradigms using human and model judges.

The paper introduces a lightweight gray-box hallucination detector combining token probabilities with vector representations of token identities.

The paper presents a multilingual benchmark to evaluate how language models resolve linter alerts while preserving program functionality.

The paper introduces a gated semantic alignment framework with multi-source rewards to improve multimodal emotional support conversations.

The paper formalizes clarification criteria to estimate model-dependent input ambiguity based on accessible and valid answers.

The paper evaluates how language exposure distribution and ordering affect the performance of multilingual models trained with limited data.

This paper surveys 138 studies on sparse autoencoders in NLP, examining dictionary design, feature recoverability, and practical utility.

This paper introduces Gap Detection, an optimization method that identifies unoccupied semantic directions to reduce homogenization in open-ended model answers.

This paper presents RM-OPD, a framework that trains domain-specialized reward models via on-policy distillation into a generalist model.

This paper introduces TreeSpark, a framework training semi-autoregressive drafters to generate candidate trees aligned with speculative verification.

This paper analyzes career path prediction systems, reporting that explicit skill representations provide marginal improvement over text-only job descriptions.

This paper proposes a joint scaling law evaluating how adding experts in multilingual mixture-of-experts models affects language capacity and loss penalties.

This paper introduces AFFIRM, using affine fields on frozen speech models to repair spoken access to recently updated text knowledge.

This paper introduces Anuvani LID, a dedicated 115M parameter model for spoken language identification across 22 Indian languages.

This paper introduces an annotated part-of-speech tagging dataset and evaluates pretrained transformer encoders for the low-resource Saraiki language.

This paper proposes PUlSE, using internal layer representations to detect factual errors and trigger step-level retrieval during medical reasoning.

This paper presents Anuvani-Diarization, evaluating zero-shot speaker diarization across 22 Indian languages with uncertainty-aware silver supervision.

This paper proposes SCOUT, a cascaded two-track framework combining visual candidate filtering, offline semantic summaries, and query refinement for document retrieval.

This paper introduces a multi-hop reasoning benchmark to evaluate fine-grained, explainable hate speech analysis in Persian.

This paper introduces FOLD, a training-free speculative decoding approach treating distant tokens as position-free far memory for long-context generation.

This paper introduces COUNT, scaling language model memory without training parameters by storing precomputed n-gram counts directly in host memory.

This paper introduces ChronoCulture to evaluate how well language models capture temporal changes in cultural attitudes across 72 countries.

This paper evaluates multilingual bias audits across ten language models, showing that lower stereotype scores often reflect translation artifacts rather than fairer behavior.

This paper presents ReaDy, a speculative decoding method using hybrid linear attention recurrent states to provide position-specific context during drafting.

Documents safety alignment methods, negative preference tuning outcomes, and automated judge limitations for a Portuguese language model.

Compresses long-context key-value caches using test-time training on generic document prompts without task-specific labels.

Integrates clinical text and deformation history into medical image registration models for adaptive radiotherapy planning.

Uses a neural-network variational wave function to model non-Abelian quantum Hall states with realistic Landau level mixing.

Proposes an optimization method to discover undisclosed or harmful generation concepts embedded within diffusion model adaptation weights.

Evaluates closed and open models on English and Bengali abstractive summarization using lexical metrics, factuality checks, and automated judges.

Improves indoor radio-map prediction by pretraining models on diverse synthetic floorplans and an open-source ray tracer.

Introduces a welfare-aware objective to allocate limited human feedback across diverse individuals when training pluralistic alignment models.

Proposes input loss curvature as a sampling-free metric to detect the onset of training data memorization in diffusion models.

Combines spectral graph coarsening with continuous relaxation to find high-quality solutions for compute-constrained MAX-CUT problems.

Uses reinforcement learning to train an auxiliary model that inspects execution traces and generates targeted prompts for another system.

Studies how the choice of local optimizers affects model performance degradation when scaling distributed training replicas under fixed budgets.

Shows that incorporating static item value into dense retrieval outperforms generative retrieval in e-commerce recommendation settings.

Introduces a prompt optimization framework that monitors item-level regressions and recoveries rather than relying solely on aggregate accuracy.

Introduces a generalist policy that tracks diverse dexterous hand-object interactions using value-guided reset allocation during training.

Surveys dataset design, construction methods, and evaluation protocols for improving spatial reasoning in multimodal language models.

Analyzes how language models develop in-context learning by separating simple demonstration copying from multi-token rule composition across training checkpoints.

Steers language model activations using vectors derived directly from the model's own training data or inferred proxy data.

The paper proposes a training-free structured pruning method that trims multilayer perceptron layers while preserving attention mechanisms for reasoning tasks.

The paper demonstrates onboard thermal segmentation training and closed-loop attitude control on a satellite in low Earth orbit.

The paper introduces a transformer framework trained on matched molecular pairs for prompt-guided, multi-property molecular optimization in low-data regimes.

The paper proposes sampling from a geometric horizon world model to imagine novel outcomes for zero-shot policy generalization.

The paper presents a system using technical language processing to transform industrial maintenance work order records into an ontology.

The paper proposes confining feature distillation loss to language-routed adapters to balance performance in compact multilingual speech recognition models.

The paper introduces a benchmark assessing language models on operational planning from incomplete, complex, and contradictory military document sets.

The paper introduces an evergreen benchmark to evaluate the ability of language models to retrieve and format multi-element factual lists.