Model releases, architecture work and prompting research on large language models — from frontier-lab announcements to the arXiv papers behind them.
arXiv:2602.10480v4 Announce Type: replace Abstract: Large language models (LLMs) exhibit strong general-purpose reasoning capabilities, yet they frequently hallucinate when used as world models (WMs)...
arXiv:2609.07595v2 Announce Type: replace-cross Abstract: The same underlying computational problem is solved across unrelated fields under different names: recursive Bayesian state estimation appear...
arXiv:2609.16664v1 Announce Type: new Abstract: Leveraging Large Vision-Language Models like CLIP has recently set new benchmarks for No-Reference Image Quality Assessment (NR-IQA). However, the cont...
arXiv:2609.16785v1 Announce Type: new Abstract: Zero-shot anomaly detection aims to localize anomalies without target-domain samples. Existing CLIP-based methods suffer from coarse anomaly maps and l...
arXiv:2609.17021v1 Announce Type: new Abstract: Autonomous wheel-loader control requires joint reasoning over task semantics, egocentric vision, proprioception, and 3D scene geometry. We present sens...
arXiv:2605.05556v2 Announce Type: replace Abstract: Artificial neural networks trained on visual tasks develop internal representations resembling those of the primate visual system, a discovery that...
arXiv:2609.15991v1 Announce Type: new Abstract: Standard subword tokenizers either treat every orthographic variation of a word (such as hello, Hello, HELLO, and H\'ello) as unrelated vocabulary entr...
The paper introduces JudgeBiasBench, a benchmark that systematically quantifies judgment biases in large language model (LLM)-based judges across four dimensions and 12 bias types. It evaluates both generative and discriminative judges, revealing significant bias patterns that undermine reliability. The authors propose bias-aware training—reinforcement learning for generative judges and contrastive learning for discriminative judges—to reduce these biases while maintaining evaluation performance.
arXiv:2609.13677v1 Announce Type: cross Abstract: Modern real application problems involve matrix-valued parameters, yet conventional optimizers treat them as vectors, thereby motivating matrix-aware...
arXiv:2609.16464v1 Announce Type: cross Abstract: The clinical management of autism spectrum disorder (ASD) faces a bottleneck in early screening, mainly because trained specialists are scarce and co...
arXiv:2609.17419v1 Announce Type: new Abstract: Long-horizon LLM agents must maintain task state across extended sequences of observations, actions, tool calls, and intermediate beliefs. We study the...
arXiv:2609.16752v1 Announce Type: new Abstract: Cognitive Field Theory (CFT) proposes that cognition arises from memory-dressed collective dynamics that generate a persistent macroscopic cognitive fi...
Multi-encoder fusion extends Large Audio-Language Models (LALMs) beyond speech-centric recognition, but selecting encoders via intuition or exhaustive search often introduces redundant representations...
Large language model optimization is an active research area, spanning quantization of model weights, early-exit methods for skipping layers, and speculative decoding. Each track uses its own quality...
Speech Large Language Models (Speech-LLMs), typically built from a pre-trained speech encoder, a modality projector, and an LLM fine-tuned with Low-Rank Adapters (LoRA), have shown strong Automatic Sp...
Emotion recognition in conversation (ERC) is a production capability behind agent-assist prompts, escalation routing, and post-call analytics in contact-center-as-a-service (CCaaS) platforms, where co...
Google has launched Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, two speech‑to‑speech models similar to OpenAI’s GPT‑Live family. A web UI built with GPT‑6 Astra Extra High lets users select a model, choose a voice preset, provide an optional system prompt, and engage in voice conversations directly in the browser, even interrupting the model while it speaks. The implementation relies on a WebSocket endpoint and the Web Audio API for capturing and playing audio, with no external libraries required.
We study the problem of demonstration selection, which involves selecting a subset of examples for prepending to a query to a language model. This problem is closely related to in-context learning and...
Large language models can produce fluent answers when their factual support is weak. This paper introduces Chain-of-Self-Questioning (CoSQ), a prompt-only framework that makes answer commitment condit...