Recent progress in large language model reasoning has been driven by benchmarks and reinforcement learning environments with automatically verifiable rewards, particularly in mathematics, code, and formal logic. These settings make model accuracy easier to evaluate and optimize, but it remains unclear how far success u...
Ibrahim Ethem Deveci, Funda Tan Çalık, Barış Deniz Sağlam et al.· 0 citations
GoldiMask is introduced, which selects tokens to reveal as context by approximately maximizing a submodular objective and weights the remaining targets according to how they benefit from the selected context and their remaining learning potential.
L. Mualem, L. Pastor-Pérez, Vinh Tong et al.· 0 citations
This work shows that training GFlowNets to generate latent conversation structure using a Gaussian mixture density over key interaction features enables sampling expert strategies in proportion to their prevalence in the training data to provide a stronger training signal than competitive synthesis baselines.
Sumit Asthana, Michael Ion, Kevyn Collins Thompson· 0 citations
Recently, Large Language Model (LLM)-based Text-to-Speech (TTS) models have achieved remarkable naturalness. However, the standard Supervised Fine-Tuning paradigm often converges to statistically averaged prosody, limiting emotional expressiveness. While preference-driven optimization offers a promising alternative, ex...
Sihang Nie, Xiaofen Xing, Rui Xing et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Reliable provenance for LLM outputs requires multi-bit watermarks that remain robust under editing while maintaining low false-positive rates. Existing ECC-based LLM watermarks rely on hard-decision decoding, discarding token-level reliability information and limiting robustness under post-generation edits. We propose...
Automated code documentation is essential for modern software development, providing the contextual grounding that both human developers and coding agents rely on to navigate large codebases. Existing repository-level approaches process components independently, causing redundant retrieval and conflicting descriptions...
Suyoung Bae, Jaehoon Lee, Changkyu Choi et al.· 0 citations
Audiovisual speech recognition (AVSR) combines acoustic and visual cues to improve transcription robustness under challenging conditions but remains out of reach for most under-resourced languages due to the lack of labeled video corpora for training. Synthetic visual data have been shown to be an effective augmentatio...
This work evaluates entity tracking in both language models and humans using naturalistic narratives at multiple levels of complexity, finding that entity tracking degrades specifically with narrative complexity, not narrative length.
We present the first systematic study of massive activations (MAs) in layer-interleaved Hybrid linear attention large language models (HLA LLMs), examining their architectural organization, training-time emergence, underlying mechanisms, and functional significance. Across five linear attention architectures, six hybri...
Zunhai Su, Bohan Sun, Xialie Zhuang et al.· 0 citations
Diversity in LLM mathematical reasoning is critical for exploration, but common diversity metrics mostly capture surface-level variation rather than differences in how a problem is solved. We address this gap by introducing approach-level diversity: variation in strategies across correct solutions to the same problem....
Sangmook Lee, Minbeom Kim, Jeonghye Kim et al.· 0 citations
Medical large language models are commonly adapted with a fixed low-rank budget, even though medical questions differ substantially in confidence, clinical coverage, and cross-domain difficulty. We study adaptive rank budgeting for parameter-efficient medical question answering: for each question, the adapter decides w...
Shucan Ji, Yining Huang, Hongliang Guo· 0 citations
IndicKLAR, an Indic extension of the KLAR-CLC benchmark covering 18 of the 22 scheduled Indian languages, is introduced and a consistent flip point is observed between the transition from incorrect to correct prediction---between the native and code-mixed settings.
Debajyoti Mazumder, Divyanshu Pathak, P. Kodali et al.· arXiv.org· 0 citations
Flint-interactive gives AI agents a language for creating interactive visualizations. Click a chart, filter data, highlight related content, and explore beyond the view the author originally chose. The post Flint-interactive: A language for creating interactive data visualizations appeared first on Microsoft Research.
