Knowledge catalogue
A month ago I posted about Gemma sometimes solving the reasoning problems inside my translation data instead of translating them. A few people suggested two very fixes, which is to use a proper transl
arXiv:2512.09108v2 Announce Type: replace-cross Abstract: Agentic AI systems built on large language models (LLMs) offer significant potential for automating complex workflows, from software developme
arXiv:2609.03611v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) are increasingly used to evaluate robot manipulation outcomes, but existing benchmarks offer limited evidence of cross-d
For how much I use AI and automation, you'd THINK I run 100% of my business through Codex or Claude Code or Slack. But I still use native apps. I talk to several people who brag that 'they don't even
for the first time in a year, @OpenAI's model launch has been better received than a Claude launch. https://www.latent.space/p/ainews-gpt-6-astra-openais-biggest this is a huge upset that we had previ
arXiv:2609.03331v1 Announce Type: new Abstract: Vision-language models (VLMs) are increasingly deployed in multi-turn settings where users may describe visual content with incorrect assumptions. Yet e
arXiv:2609.03370v1 Announce Type: new Abstract: In frame semantics, sentence comprehension is assumed to proceed by relating lexical meaning to background knowledge called semantic frames, thereby ena
arXiv:2609.03807v1 Announce Type: cross Abstract: A free pause token gives a language model extra compute to form each next-token prediction (as a pause, or thinking, token does) but carries that comp
arXiv:2609.03350v1 Announce Type: cross Abstract: Pretraining data for Armenian, a morphologically rich and low-resource language, is scarce, and no open Armenian LLM has been released with the data a
Gemini 3.8 Flash might be maxxed on other benchmarks, but not quite on document parsing. We benchmarked it on ParseBench. It does slightly better on tables + content faithfulness compared to Gemini 3.
arXiv:2609.03193v1 Announce Type: cross Abstract: Nested sampling (NS) resolves the thermodynamics of an atomistic system from a single simulation, but its practical reach is limited by the Markov-cha
arXiv:2606.12821v2 Announce Type: replace Abstract: Environmental scientists spend disproportionate effort on data wrangling rather than analysis. New AI agents can be a helpful tool, but no benchmark
arXiv:2609.03553v1 Announce Type: new Abstract: Policy analysis requires more than predicting whether a proposal will pass: it requires identifying who will be affected, how those actors respond, and
arXiv:2609.03361v1 Announce Type: cross Abstract: We propose a geometric parametrization of the filters in a single convolutional layer: the parameter is no longer an ordered family of filter vectors,
Great tips on working with reasoning models. Normally you would append what the model figured out after the document and ask again. It turns out that where you put the reasoning trace changes long-con
arXiv:2609.03580v1 Announce Type: new Abstract: The growing scale of academic peer review has motivated the use of Large Language Models (LLMs) as review assistants, yet LLMs can generate fluent but u
arXiv:2609.04105v1 Announce Type: new Abstract: Blackwell's 4-bit floating-point (FP4) tensor cores do not automatically make attention faster because softmax conversion and on-chip dependencies domin
How good/bad is it against comparable MoEs the same size? How does it compare against Qwen 3.6 35BA3B? Since we don't have 3.8 35B this seems like an upgrade if we look at some benchmarks like termina
arXiv:2601.10187v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have achieved remarkable strides in multilingual translation but are hindered by a systemic cross-lingual verbosi
arXiv:2609.03595v1 Announce Type: cross Abstract: We investigate what makes synthetic OCR supervision transfer to real Thai documents and use the resulting insights to build Wayu-Paxa-OCR-Zero, a Thai
How is AI really affecting your job? @Cohere_Labs released the Agentic Task Ecosystem, a new dataset of 690K+ tools built for AI agents. Under a test of whether a tool could independently complete an
We all want more tokens per second but I keep seeing confusion on what to expect for given hardware. For the decoding phase (TG/s) to produce one token all the model weights and KV cache needs to be r
Hugging Face has been acquired by NVIDIA It is quite exciting to be a part of this journey! NVIDIA has been an active supporter of the llama.cpp project. For more than a year now, their engineers have
After Qwen3.8 27B came out, I decided to benchmark the models that could fit in my GPU (RTX 5080) on my actual code (C code), the results were not completely unexpected but some quants were definitely
This epyc server I am using twelve cards with 64 GB memory, plus 256GB ram. Looking at the most capable models in open source, GLM 5.3 seems to be the only option, but with Astra releasing it will lik
arXiv:2609.03052v1 Announce Type: new Abstract: As services move online, trust institutions such as banks, lenders, and governments must verify the identity of remote users. Fraud detection tools are
arXiv:2609.03781v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used in multilingual settings, yet their safety is still evaluated primarily in English. This limits our
arXiv:2609.03422v1 Announce Type: new Abstract: Diversity is a widely observed factor in the resilient function of collective systems, yet the type of diversity that matters depends on the properties
arXiv:2609.03966v1 Announce Type: new Abstract: Agent evaluations report a tool-call rate read off the serving stack. That number can be zero while the model is emitting well-formed calls: the interfa
arXiv:2609.03329v1 Announce Type: cross Abstract: Long-duration gravitational-wave modelling must resolve fast orbital motion together with slow dissipative evolution while preventing small numerical
arXiv:2609.03967v1 Announce Type: cross Abstract: Several studies have evaluated the ability of Large Language Models (LLMs) for meal planning, yielding positive outcomes. These models can process nat
arXiv:2609.04030v1 Announce Type: new Abstract: IRWOZ has improved industrial human-robot interaction (HRI) dialogue systems through domain-specific annotations. However, its initial version contains
It happened again... this time OpenAI's rogue agents cyber-attacked (well, spammed) a dormant German wiki and used it to share the answers to a benchmark they were training against https://simonwillis
It is funny that the Fermat's Last Theorem proof description, short as it is, still smells so much of Claude ('names each step and the Lean Theorem that carries it'). Anthropic just uploaded a Lean 4
arXiv:2609.03436v1 Announce Type: cross Abstract: Reasoning traces of large language models are widely read as containing 'breakthrough' moments and early-legible fates. Both readings rest on measurem
arXiv:2609.03181v1 Announce Type: new Abstract: We present Jina-OCR-v1, an end-to-end document parsing model built to serve on low-budget GPUs. It combines the compressed-vision encoder and the 3B mix
arXiv:2609.03588v1 Announce Type: new Abstract: As LLMs increasingly act through tools, they must reconcile user instructions, parametric knowledge, and dynamic environmental observations before takin
arXiv:2603.12440v2 Announce Type: replace-cross Abstract: GPU kernel optimization challenges LLMs beyond standard coding tasks, as it requires an understanding of hardware architecture, parallel compu
arXiv:2609.03597v1 Announce Type: new Abstract: Land ownership in Bangladesh is recorded in Ana-Ganda-Kora-Kranti-Til, a base-16 positional fraction system with dedicated Unicode glyphs, no mainstream
arXiv:2609.04173v1 Announce Type: new Abstract: For scientific progress, we need benchmarks that test the limits of state-of-the-art models, and evaluation methods that inform us about failure cases.
arXiv:2609.03294v1 Announce Type: new Abstract: Latent world models support efficient model predictive control from high-dimensional observations, yet optimizing a single learned latent objective can
arXiv:2503.10421v2 Announce Type: replace Abstract: The application of learning based methods to vehicle routing problems has emerged as a pivotal area of research in combinatorial optimization. These
arXiv:2606.20954v2 Announce Type: replace-cross Abstract: Long-running language-model systems accumulate interaction history that outgrows the context window, so they must continually evict. When an e
arXiv:2609.03594v1 Announce Type: cross Abstract: As integrated circuit technology scales into the nanometer regime, the traditional disconnect between logic synthesis and physical design has led to s
arXiv:2609.02954v1 Announce Type: new Abstract: Identifying the issues disputed between litigating parties is a crucial component of real-world litigation. However, legal issues remain comparatively u
arXiv:2505.00284v3 Announce Type: replace-cross Abstract: Rapid advances in vision-language models (VLMs) have generated growing interest in their application to autonomous driving. A prevailing assum
arXiv:2609.03796v1 Announce Type: cross Abstract: We introduce LLaDA-Image, a unified framework that pairs a 6B Diffusion Transformer (DiT) trained from scratch with a frozen vision-language understan
arXiv:2609.02931v1 Announce Type: cross Abstract: Scripted and rule-based non-player characters (NPCs) in combat video games often exhibit predictable behaviors that experienced players can exploit, w
arXiv:2606.18388v2 Announce Type: replace-cross Abstract: RL post-training strategies are dataset-dependent and reveal a recurring empirical pattern: capacity parameters accumulate monotonically acros
arXiv:2609.03426v1 Announce Type: new Abstract: Transformers lack a native lookup mechanism, requiring repeated dense computation to recognize and reuse local static patterns. Lngram v1 introduces tok
arXiv:2609.03225v1 Announce Type: new Abstract: End-to-end autonomous driving has increasingly adopted world model-based reinforcement learning frameworks to improve learning efficiency through extit{
arXiv:2609.03507v1 Announce Type: cross Abstract: Tracking depression from multi-session counseling dialogues requires estimating both current symptom severity and how it changes across sessions. Yet
arXiv:2609.03511v1 Announce Type: new Abstract: Large Language Models (LLMs) demonstrate strong multilingual reasoning performance, yet their robustness to semantics-preserving structural variation re
arXiv:2603.18261v2 Announce Type: replace-cross Abstract: Neural Representations for Videos (NeRV) encode entire video sequences within neural network parameters, offering an alternative paradigm to c
arXiv:2609.03493v1 Announce Type: new Abstract: Modern vision-language models (VLMs) can directly answer many image-grounded questions, yet they often struggle with complex queries requiring fine-grai
arXiv:2609.02897v1 Announce Type: new Abstract: Speculative decoding accelerates LLM inference by drafting candidate tokens and verifying them in parallel. Tree-attention drafters such as EAGLE-3 are
arXiv:2508.00935v3 Announce Type: replace-cross Abstract: Computer-using agents (CUAs), which can autonomously control computers to perform multi-step actions, might pose significant safety risks if m
arXiv:2609.03261v1 Announce Type: cross Abstract: A benchmark score credits final answers, but not the route by which an item can be answered. In medical multimodal multiple-choice questions (MCQs), t
arXiv:2609.03690v1 Announce Type: new Abstract: The joint interpretation of metabolic function and anatomical structure is essential for clinical diagnosis in whole-body PET/CT. Although recent advanc
arXiv:2609.03464v1 Announce Type: new Abstract: Personally Identifiable Information (PII) detection is a foundational component of data protection infrastructure where missed entities constitute direc