Model releases, architecture work and prompting research on large language models — from frontier-lab announcements to the arXiv papers behind them.
arXiv:2609.36692v1 Announce Type: cross Abstract: Matrix optimizers have emerged as a promising direction, with Muon standing out as a prominent design. Revisiting Muon through its full-Gram represen...
arXiv:2609.36794v1 Announce Type: new Abstract: Signal features derived from photoplethysmography (PPG) require different signal durations to characterize. Existing PPG foundation models treat durati...
arXiv:2609.36813v1 Announce Type: new Abstract: Large language models (LLMs) exhibit strong general capabilities that mechanistic interpretability has attributed to sparse computational circuits. How...
arXiv:2609.36890v1 Announce Type: new Abstract: In scientific machine learning, physical fields governed by partial differential equations exhibit low-rank structure and scale invariance. When solvin...
arXiv:2609.37750v1 Announce Type: cross Abstract: Pulmonary embolism (PE) is a leading cause of cardiovascular mortality, yet the real-world performance of FDA-cleared AI detection models remains inc...
arXiv:2609.37818v1 Announce Type: cross Abstract: Empathetic spoken dialogue requires models to use both what is said and how it is said to decide how to respond. Explicit CoT can improve paralinguis...
arXiv:2609.35599v2 Announce Type: replace Abstract: When recalling lists of concepts (e.g., animals) during the semantic fluency task (SFT), both humans and large language models (LLMs) organise thei...
Vision‑Language Models (VLMs) used in autonomous driving are evaluated on their ability to understand velocity through three tasks: surrounding‑agent speed, current ego speed, and short‑horizon future ego‑speed. Experiments on nuScenes show that while current ego speed can be accessed internally, verbal outputs are poor and surrounding‑agent speed is weakly encoded; temporal cues and frame order are largely ignored. Specialized driving models (Alpamayo‑1.5) improve latent speed representations but still leave gaps between internal knowledge and verbal readout, indicating that good planning outputs do not guarantee reliable dynamic state recovery.
EviViT is a lightweight attachment for pretrained vision transformers that learns where to focus detail in high‑resolution images. It uses human visual‑search traces to supervise a question‑conditioned evidence density, guiding regional re‑reading and efficient visual token allocation. The method connects regional features to the global scene via a sparse, coordinate‑aware bridge, improving fine‑grained accuracy across nine host models while using fewer tokens than global‑only processing.
arXiv:2609.37066v1 Announce Type: cross Abstract: Post-training is central to mathematical reasoning in modern large language models (LLMs), but endpoint pass@1 alone underidentifies what has changed...
arXiv:2609.37852v1 Announce Type: new Abstract: Reliable FP8 attention remains a barrier to fully native 8-bit large language model training. We derive how forward-backward inconsistencies produce st...
The paper investigates how large language models encode and use relational information among tokens across transformer layers. By analyzing activations from prompts that require inferring relationships among three cyclic tokens (months, hours, weekdays, musical notes), the authors find a consistent layerwise progression: intermediate layers capture pairwise relationships, while later layers encode the full three‑token relationship to predict the next token. They also identify geometrically structured token relationships that do not influence prediction, and show that constraining models to use only causally relevant joint representations improves next‑token accuracy.
The paper benchmarks vision‑language models (VLMs) on two key tasks in connectomics: synapse detection (presence and polarity) and proofreading (split and merge errors). It evaluates 21 models (19 open, 2 closed) under zero‑shot, few‑shot, and LoRA fine‑tuning, comparing them to specialist models on datasets built from public resources. While most VLMs perform at chance in zero‑shot settings, LoRA fine‑tuning with a few thousand labels brings open models to specialist performance, and adapted VLMs outperform specialists on unseen species for merge‑error detection.
The paper introduces a controlled evaluation to disentangle answer coverage, repeatable task advantages, and gains from pre‑execution selection in large‑language‑model (LLM) harnesses. On 386 MATH‑500 tasks, eight generated harnesses and a baseline with nine identical copies were compared over three executions each, revealing that identical programs provide a 2.16‑point repeat‑averaged oracle headroom while generated programs show more repeatable score patterns but mainly expose persistent weaknesses. The study concludes that coverage and repeatability alone cannot justify claims of useful specialization and proposes an evaluation standard for harness diversity that requires task advantages to persist across executions and improve on additional fixed‑program executions under matched inference budgets.
The study evaluates multi‑agent debate (MAD) in small language models, testing whether cognitive diversity—via personas, sampling temperature, or model identity—drives performance gains. Across 23 models, five tasks, and over 5,500 runs, MAD consistently outperforms single‑model inference but, when matched for generation budget, it ties or falls behind self‑consistency sampling, with persona prompting actually reducing accuracy. The authors find that MAD’s benefits largely stem from the first answer exchange and that many reported gains are due to ensemble‑sampling effects rather than true diversity, highlighting the need for budget‑matched, contamination‑checked baselines. whyItMatters":"The findings clarify that MAD’s perceived advantages may be overestimated and that future debate mechanisms must be evaluated against rigorous, budget‑matched baselines to ensure genuine performance improvements."
The paper investigates how scaling a team of small language‑model agents affects performance across different orchestration architectures. By testing eight architectures on five short‑answer benchmarks and an executable‑code benchmark, it finds that team scaling yields large gains on arithmetic word‑problem tasks but only modest improvements on multiple‑choice and code generation tasks, with no single architecture dominating all tasks. The authors explain these patterns using a generate‑transform decomposition that separates coverage and transformation effects, showing that arithmetic tasks benefit from both coverage and critic‑guided transformation, while other tasks are limited by saturation or poor conversion.
The paper introduces the Distributed‑Evidence Paradox, where long‑running LLM agents compress past interactions into persistent memories that may not be fully supported by the interaction history. It defines three key requirements—evidence scope, compositional validity, and admission reliability—and proposes DerivAudit, a framework that checks whether a memory is truly supported by the available history. Experiments on two memory corpora show that expanding the evidence base can recover support for many memories, yet many remain unsupported, and broader evidence alone does not guarantee reliable admission.
The paper introduces REST (REpresentation‑Supervised Thoughts), a new training objective for latent recursive language‑model systems that supplements cross‑entropy loss with differentiable penalties enforcing causality, minimality, separability, and stability of internal thought representations. By applying REST to both single‑agent and multi‑agent setups without changing architectures or adding inference parameters, the authors achieve up to 7.5 percentage‑point gains in accuracy across seven diverse benchmarks and a 30 % improvement in convergence to the final answer. The method also yields more informative latent thoughts, improving interpretability of agent communication.
The paper introduces OTROPE, a likelihood‑free method for off‑policy evaluation of large language models (LLMs) that uses optimal transport to align labeled samples from a behavior model with unlabeled samples from a target model in a semantic space. OTROPE corrects human‑labeled residuals with proxy predictors, achieving a doubly robust evaluation without requiring behavior‑policy modeling or density‑ratio estimation. The authors provide theoretical guarantees for consistency and convergence, and demonstrate through synthetic and real LLM tasks that OTROPE outperforms existing baselines and can elevate weaker evaluators to match or exceed stronger ones.
The study examines how four large language models (Gemma-3-4b-it, Qwen2.5-7B-Instruct, Llama-3.1-8B-Instruct, and GPT-5-nano) exhibit repetition effects in a social media simulation. Using a two‑phase within‑context design, researchers collected 336,000 ratings on truth, importance, sentiment, and interest for 100 statements across 10 feed variants and 3 replications. Results show distinct patterns: Gemma-3 displays a genuine Illusory Truth Effect, Qwen2.5 shows a mere exposure effect, GPT-5-nano shows no truth boost and mild skepticism, and Llama-3.1 shows a small truth boost with reduced evaluative dimensions.