Knowledge catalogue
Hello everyone! A little while back I posted about LlamAmpere, a fork of Llama.cpp with Ampere-specific improvements (though it is caught up to main and will support other hardware, too). Thank you to
arXiv:2609.31224v1 Announce Type: cross Abstract: Full-duplex speech language models continuously accumulate acoustic key-value (KV) states, making long-running interactions memory-intensive. During l
Security operations teams are changing their approach in the world of artificial intelligence, adopting automation to turn signals into outcomes while maintaining the right context and control. This t
arXiv:2609.31168v1 Announce Type: new Abstract: My doctoral work aims to explore Audio Emotion Recognition (AER) in the context of atypical listening. This research focuses on auditory hypersensitivit
arXiv:2609.30293v1 Announce Type: cross Abstract: The Model Context Protocol (MCP) enables AI agents to discover and call tools, but loading every definition becomes expensive as connected catalogs gr
arXiv:2604.04247v2 Announce Type: replace Abstract: Recent advances in prompt learning allow large language model agents to acquire task-relevant knowledge from inference-time context without paramete
arXiv:2609.30977v1 Announce Type: cross Abstract: Uniform discrete diffusion models (UDMs) commonly use explicit time conditioning, but we find that it can often be unnecessary in practice. In this pa
arXiv:2609.30818v1 Announce Type: cross Abstract: Multi-modal planning is promising for autonomous driving by representing multiple plausible behaviors in ambiguous and long-tail scenarios. Existing m
arXiv:2609.30954v1 Announce Type: new Abstract: Large language models operating in multilingual contexts must resolve target response languages early in generation, yet the causal circuitry governing
xAI's Grok 4.7 is now available on Amazon Bedrock: a frontier model for coding, long-running agents, and knowledge work. It offers a 500K token context window and four configurable reasoning effort le
arXiv:2504.08596v3 Announce Type: replace-cross Abstract: Hallucination, the generation of non factual content by AI systems, poses serious risks in medical contexts, where errors can directly affect
arXiv:2609.31255v1 Announce Type: new Abstract: General-purpose agent memory summarizes conversations: it extracts salient snippets, embeds them, and retrieves the top-k into the prompt. A health agen
arXiv:2609.30554v1 Announce Type: new Abstract: Large language models (LLMs) have recently demonstrated promising capabilities as in-context policy optimizers for Reinforcement Learning (RL), enabling
arXiv:2504.06188v3 Announce Type: replace Abstract: AI agents can extend their capabilities at inference time by loading reusable skills into context, yet equipping an agent with too many skills, part
arXiv:2609.30560v1 Announce Type: new Abstract: Beyond collision avoidance, socially competent robot navigation requires adherence to implicit social conventions that vary across contexts, cultures, a
Hi! I'm the author of MoEspresso, which is my way of putting my own ideas about inference engines to the test. A lot of the fun has been trying different design choices, measuring what happens, and fi
We built an inference engine for MoE models that don't fit in VRAM + RAM. Most of the model stays on the SSD, and experts are read as tokens need them. This started as a proof of concept, and poc work
Revert 'Change max context length for auto-fitting with unified KV (#28849)' (#29437) This reverts commit b04d4e5. Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attest
I set out to test portable Engrams and accidentally ended up testing compiled external memory instead. Am I onto something useful or reinventing a known idea? I've been building a small open research
arXiv:2609.28609v1 Announce Type: new Abstract: Role-playing agents based on large language models have been widely applied in areas such as personalized assistance and social simulation. Recent RL me
Runtime governance is becoming critical as autonomous agents move from answering questions to taking action across enterprise systems. Organizations need to govern both the business context agents rec
arXiv:2609.29879v1 Announce Type: new Abstract: Generative modeling approaches often focus on recovering broad statistical characteristics from the training data. In the context of graph generation, t
arXiv:2609.30147v1 Announce Type: new Abstract: Large Language Models (LLMs) typically exhibit a performance profile where reliability degrades as task complexity increases. We address the challenge o
I had Mica (my 4B decision model) and Kev 4B play the same Tetris games, same seed and same piece order, one RTX 3090. For context, this is a side project. Training and all the experiments ran on rent
arXiv:2609.28877v1 Announce Type: cross Abstract: New words are invented every day. A human listener can learn a new word by hearing it clearly once and inferring its usage from sentence context. This
arXiv:2609.28605v1 Announce Type: cross Abstract: The Factivity Inference Evaluation 2026 (FIE2026) classifies Chinese context-hypothesis pairs into nine ordered factivity intervals. Its evaluation me
arXiv:2609.29555v1 Announce Type: new Abstract: Navigation World Models (NWMs) predict action-conditioned visual futures for planning. Two practical challenges are central to their design: selecting a
arXiv:2609.28870v1 Announce Type: cross Abstract: Long-running LLM applications repeatedly send growing context, making prefix caching critical for reducing prefill cost. Yet prefix-cache behavior und
Learn how Aderant built an intelligent ticket triage system on Amazon Nova Lite through Amazon Bedrock, automating context gathering, classification, routing, and knowledge enrichment for its cloud op
arXiv:2609.27373v1 Announce Type: new Abstract: Recurrent language models repeatedly apply shared network blocks to refine latent representations, but standard inference recomputes global attention at
arXiv:2609.27208v1 Announce Type: new Abstract: Spatial transcriptomics (ST) measures gene expression in tissue context, but dense capture grids can be costly and may repeatedly sample morphologically
arXiv:2609.27327v1 Announce Type: new Abstract: Video provides a rich record of human behavior, interaction, and situated contexts, offering important evidence for understanding people and conducting
arXiv:2609.27470v1 Announce Type: cross Abstract: Recent video-language models increasingly adopt hybrid architectures that interleave linear and full attention layers for efficient long-context proce
arXiv:2609.27233v1 Announce Type: new Abstract: Transformer language models process sequences token by token in an autoregressive manner, making growing contexts increasingly expensive. Yet many adjac
Built this because I got tired of guessing whether a model would fit in my 6GB VRAM before downloading it, and because ollama.com's library doesn't let you sort by size. What it does: - Benchmarks Oll
arXiv:2609.28256v1 Announce Type: cross Abstract: Vision-Language-Action models provide a strong foundation for general-purpose robot control, yet a vast majority of policies do not preserve and lever
arXiv:2609.27349v1 Announce Type: new Abstract: Real-world molecular design remains challenging for large language model (LLM)-based agents. It requires them to interpret design contexts, satisfy mult
arXiv:2609.28122v1 Announce Type: cross Abstract: Neural Processes (NPs) are model-based meta-learners that implicitly learn a stochastic process and adapt to a new task from a small context set. Most
Many a praise have been sung on Qwen-3.8, but here is mine. Qwen-3.8 and I had a rocky start, because it thinks so much. Watching it working is painful, so you have to stop doing that. You have to let
A while ago I posted 15 tok/s output and 100-120 tok/s prompt processing with the IQ3_XXS quant on a 12GB RTX 5070 using llama.cpp. Since then I built my own inference engine for this one model and th
arXiv:2609.27612v1 Announce Type: new Abstract: Long-horizon robot execution requires a clear distinction between a model's proposal, a controller's termination, and verified task completion. We prese
Knowledge management startup Stravito AB today introduced a Model Context Protocol server that lets people access the company’s consumer and market research from AI tools such as OpenAI Group PBC’s Ch
arXiv:2609.28429v1 Announce Type: new Abstract: An always-on robot faces an endless stream that never resets: instructions arrive and lapse, the scene changes, and its own past actions reshape what it
In real-time streaming, every millisecond counts. For example, at 30 frames per second (fps), developers have a strict frame budget of just 33.3 ms (and only 16.6 ms at 60 fps) to ingest camera frames
arXiv:2609.25541v1 Announce Type: new Abstract: Tabular foundation models learn to predict cell values in context, whereas world-model self-supervision asks for prediction in representation space (LeC
arXiv:2609.25939v1 Announce Type: new Abstract: The performance of large language model (LLM) workflows often depends on selecting a small set of in-context demonstrations to guide model behavior on n
HEMA, a 100-year-old Dutch retailer, turned developer portal-hopping into instant answers by building HAL, an internal AI assistant on Amazon Bedrock AgentCore. Using Model Context Protocol (MCP), HAL
Hi, this is the research project I'm working on. I must say, building a GUI harness is way more complicated than I thought. The goal is something like Minecraft. You know, a few simple tools that a us
arXiv:2609.25611v1 Announce Type: new Abstract: We introduce Qwen3.8-Omni-Flash, a natively multimodal agentic model for real-world multimodal productivity. Compared with previous omni models, which p
arXiv:2609.24710v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly used in cybersecurity, where accurate analysis often requires multi-step and context-dependent reasoning
arXiv:2609.26304v1 Announce Type: new Abstract: Tendon-driven continuum manipulators are widely used in medical applications, where accurate tip-position estimation is essential for precise navigation
arXiv:2603.05294v3 Announce Type: replace Abstract: Existing LLM-based web agents struggle on complex, long-horizon tasks due to limited in-context memory, weak planning abilities, and greedy behavior
Overview This release focuses on backend performance and correctness, broader model coverage, and more robust server/router operation. It adds HRM-Text (DFM Mimir 1B) support, MiMo-V2.6 and HunyuanOCR
Saw a new open 27B model release that looks like it could be interesting for the Ollama crowd: VeriLoop E2. It is a Qwen3.8-27B post-train aimed at code, mathematics, scientific reasoning, and agentic
arXiv:2609.25709v1 Announce Type: new Abstract: Environmental reconnaissance needs mobile platforms that carry sensors, preserve measurement context, and return interpretable evidence under limited en
arXiv:2609.22796v1 Announce Type: new Abstract: Dialectal Arabic machine translation (MT) remains challenging despite recent progress in Arabic language technologies, particularly because effective tr
arXiv:2609.24059v1 Announce Type: new Abstract: Semantically meaningful subtask labels can provide useful contexts for long-horizon policies, but automatically identifying both reliable temporal bound
arXiv:2609.22977v1 Announce Type: cross Abstract: Similarity-based retrieval is the dominant rule for conditioning large language models (LLMs) in in-context learning, retrieval-augmented generation,
arXiv:2606.20002v2 Announce Type: replace-cross Abstract: This work presents a general framework for training large language models (LLMs) to 'Connect the Dots' (CoD), a meta-capability required by lo
arXiv:2609.22641v1 Announce Type: new Abstract: Autoregressive video world models enable temporally coherent generation for a single observer. Extending them to multiple agents requires consistency ac