Natural Language Processing

Anthropic's Claude Code processes code through a three-stage pipeline of parsing, pattern matching, and token-by-token generation, rather than executing the code directly. The AI assistant breaks prompts into tokens—aver…

A product manager at a 60-person SaaS company who used Notion AI daily for a year found that the tool saves significant time on meeting note compression and standup generation but produces dangerously inaccurate results …

Researchers have developed ForestHG-Trace, a framework enabling traceable long-horizon ecological reasoning over large-scale forest scenes by representing multimodal NEON forest data as ecological hypergraphs. The system…

Researchers developed a translation model that converts soft prompts—parameter-efficient but uninterpretable LLM adaptations—into natural language text. The translated prompts outperformed existing interpretability metho…

Researchers introduced MTM-Bench, a controlled benchmark that isolates three distinct language roles—instruction, content, and response—across English, Spanish, and Chinese to evaluate multilingual LLM task execution. Te…

Researchers at Simorgh have developed a region-aware hybrid retrieval method combining BM25 lexical matching and dense semantic similarity with regional weighting heuristics to improve culturally grounded question answer…

Researchers have developed a new interpretation method for Transformer models that use heterogenous attention structures, which process information from different sources. The approach addresses challenges in understandi…

Researchers evaluating English-to-Hindi translation systems found that five major generative models frequently erase gender cues through ergative and honorific constructions, failing to preserve gender recoverability fro…

Researchers have found that hallucinations in small language models (SLMs) can be useful for solving complex multi-step reasoning problems. A new study proposes a "answer first-reason later" framework that allows SLMs to…

A new study published on arXiv reveals that language models consistently learn to verify factual knowledge before they can generate it, creating a "generation-verification gap" that persists across training phases. Resea…

Researchers have developed StoryMI, a multi-agent framework using large language models to generate controlled motivational interviewing dialogues for psychotherapy. The system uses questionnaire-based client profiles ex…

Researchers have developed a new Large Language Model (LLM)-based architecture that detects and quantifies human values in text, addressing limitations of previous approaches tied to specific value theories or complex pr…

Researchers have developed EvoSpec, a framework that enables real-time evolution of draft models in speculative decoding through dynamic vocabulary and parameter adaptation. The system addresses the failure of static pru…

Researchers have developed BioELX, a two-stage cross-lingual biomedical entity linking framework that requires no task-specific annotated training data. The system improves candidate retrieval by enriching SapBERT with m…

Researchers introduced CARE (Community-Aware Reaction Evaluation), a human-AI collaborative framework that benchmarks large language models against authentic community responses to real-world news by analyzing fine-grain…

Researchers at the University of Cambridge have identified a fundamental trade-off between stability and expressivity in spoken language models trained on synthetic data, where phonetic accuracy improves but prosodic var…

Researchers have developed OralAgent, the first dental-specialized AI agent that integrates 22 visual analysis tools and 368 dental textbooks to autonomously perform multimodal reasoning, tool-based decision-making, and …

Researchers have developed RAG-Coding, a method that uses four large language model agents to automate ICD-10-CM medical coding by grounding decisions in official coding tabular lists and guidelines. On the MDACE dataset…

Researchers have developed techniques to enable fine-grained and within-utterance speaking style control in prompt-based text-to-speech models. The methods allow smooth style interpolation between utterances and time-var…

Researchers have developed DeepSciVerify, a two-stage pipeline that verifies scientific claim-citation alignment by first analyzing abstracts and escalating uncertain cases to full-text evidence only when needed. The sys…