Search: “context”

I've been using Claude Opus 5.5 to speed up Qwen3.8-27B on my PC (rtx 3090), it wrote a CUDA megakernel that is 1.4-1.9x faster than llama.cpp depending on the task/context length. The results and cod

Noah Smith (@noahopinion) has become deeply intellectually dishonest. If he actually followed the field he would a. understand the difference between pure LLMs and LLMs with neurosymbolic harnesses, c

I need you guys to understand just how insanely good Meta Muse is at idea generation. Look at your “ideas” tab. The AI labs - OpenAI, Anthropic, and Google - were always showing the same three use cas

For what it's worth. I'm a mechanical engineer by education and software engineer in practice over the past ~20 years. Playing around with local models has been a recent hobby. Today I decided to do a

Basalt: Blackwell inference engine Hi all! So over the last week I've been working on a Strata fork that's heavily tuned and can achieve throughput up to 2.6x what Strata usually does on the same weig

I stumbled across this as I was currently having glm5.3 flash run only around 10tok/s basically unusable. With Maya i am running 30 tok/s now. GLM5.3 feels even with this low quant like a much more en

No more excuses for GPU-poor folks not to start LLMing! Full 27B-class reasoning in ternary transformer weights, for llama.cpp (CUDA, Metal, CPU) ~9.3x smaller than FP16 (ideal) | 98.2% of FP16 intell

Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less 'two groups of people speaking past each other' and more a sharp funnel. - Nap

This table to help anyone looking to build a budget Data Center homelab. I copied the bulk of value based, mid level, decent speed results GPUs and feed it to AI or SI and here are the recommended res

arXiv:2609.35308v2 Announce Type: replace Abstract: Large language models treat conversation history as unverified context, so false premises injected into prior turns can be adopted as fact, a failur

Today, we announced the new Gemini agent, your single, universal agent for work. For small and midsize businesses, with lean teams and margins, Gemini will be a critical tool for you to operate effici

Alteryx One Live Query and Google Cloud BigQuery redefine how enterprises handle complex, unstructured data at scale by reducing reliance on disconnected tools, minimizing data movement, and enabling

arXiv:2610.12427v1 Announce Type: cross Abstract: Streaming Video Large Language Models (VLMs) enable continuous video understanding, yet existing benchmarks focus on low-dynamic scenarios. Under boun

arXiv:2610.10868v1 Announce Type: cross Abstract: We introduce Conversational Voice Aesthetic Model, a speech large language model for describing the voice aesthetics of real or synthetic speech respo

arXiv:2610.11542v1 Announce Type: cross Abstract: Multi-agent LLM systems negotiating with a stateful counterpart waste model calls in three ways: polite loops that never meet the counterpart's hidden

You ship an agent. It mostly works. Then the feedback loop becomes a support ticket: a user got the wrong answer, a tool call looped four times, a refund went out that shouldn’t have. Standard observa

It's been 4 years since c.ai first hallucination model, and yet, we're nowhere good enough at LLMs in terms of spontaneity/interesting hallucination features. Fortunately, last year brought new archit

arXiv:2610.13325v1 Announce Type: cross Abstract: Agentic workloads resend long contexts that barely change between calls. Prefix caching now skips most of their prefill, yet every request must still

I've been tuning llama.cpp on my Strix Halo box (Radeon 8060S, 128 GB) and the numbers are good enough to share. It's a branch called rdna35-perf on top of halo-box/strix-llama.cpp, a llama.cpp fork a

Write-up, serve command and raw data: https://github.com/lukolszewski/vllm This shows vLLM fixes and setup that let you put PLE on SSD and save 108GB of system RAM, use int8 kv and squeeze more users

arXiv:2610.13411v1 Announce Type: cross Abstract: Traditional system identification typically involves two steps: initially, training data is processed to estimate a model of the system of interest; t

arXiv:2610.13082v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly being explored as facilitators of group discussion and collaborative work, but the behavioral differences

llama : report compute graph reallocations during inference (#30329) Add an optional ggml_backend_sched_realloc_callback that user code registers with ggml_backend_sched_set_realloc_callback to be not

hexagon: add IM2COL_3D (#30305) hexagon: add IM2COL_3D support Assisted-by: OpenCode hexagon: add DMA paths for IM2COL_3D Assisted-by: OpenCode hexagon: avoid scalar fp16 conversions in IM2COL Assiste

arXiv:2610.03077v1 Announce Type: new Abstract: Propaganda detection is an essential task in natural language processing (NLP), particularly in the context of manipulative political communications. Ho

arXiv:2610.02736v1 Announce Type: cross Abstract: A compressor can keep the facts of a dialogue and still drop the turn that changed them. A user corrects a price, reverses a choice, or adds a constra

Artificial Analysis: Artificial Analysis says Mistral Large 4 is the most intelligent model from outside the US and China, achieving results comparable to DeepSeek V4.1 Flash (max) — Mistral has relea

arXiv:2610.07643v1 Announce Type: cross Abstract: Most KV-cache eviction methods ask, in effect, which memory appeared important while reading the prompt? We instead ask, which memory will matter whil

Mistral AI introduced **Mistral Large 4** in October 2026 as a general‑purpose agentic model capable of gathering information and producing final deliverables across complex workflows. The launch high

arXiv:2610.09624v1 Announce Type: cross Abstract: Tool calling, invoking external tools on demand, is central to agentic LLMs, yet the mechanism that decides whether a model calls a tool or responds d

arXiv:2607.16204v2 Announce Type: replace Abstract: Recent growth in reinforcement learning (RL) has surfaced a need for diverse, specialized training environments. Hand-curated environments with fixe

It's been 4 years since c.ai first hallucination model, and yet, we're nowhere good enough at LLMs in terms of spontaneity/interesting hallucination features. Fortunately, last year brought new archit

Automated lint: 19 errors, 64915 warnings, 3 info

Grok 4.7 is now available on Amazon Bedrock and on Google's Gemini Enterprise Agent Platform. Both platforms already offered Grok 4.6. Grok 4.7 is SpaceXAI's model for coding and knowledge work, and i

arXiv:2610.06657v1 Announce Type: new Abstract: Workplace sensing studies combine long-running behaviour traces with self-reports, yet the tools that collect those data often sit apart from the interf

arXiv:2610.06862v1 Announce Type: new Abstract: **Context:** Deep Neural Networks (DNNs) increasingly control Cyber-Physical Systems (CPSs), yet small input perturbations can cause unsafe system-level

Grace Kay / The Information: Elon Musk says Grok Bot will use the “best back end model for any given task, including Claude Opus 5.5, MidJourney, Suno” — Elon Musk announced on Tuesday night that Spac

Agent-to-agent communication with a personal agent on top is wild. Wow. I underestimated how well models like Opus 5.5 can coordinate already. This opens a whole new set of experiences with agents. Bu

Aisha Malik / TechCrunch: AI hardware startup Natura unveils Interface, a 99 smart ring that connects to AI agents and apps, including Claude, ChatGPT, Muse, Grok Bot, and Instinct — Meet Interface, a

arXiv:2610.12703v1 Announce Type: cross Abstract: High-risk applications of deep learning require precise characterization of model decision-making. We conduct a systematic explainability evaluation a

Running HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive Q4_K_M with llama.cpp at ~600 tok/s prefill and 23 tok/s decode, 131k context window, Q8 KV cache - on an RTX 2060 6GB + 32GB DDR4 RAM.

Based on my experiments and feedback from claude, it seems extremely unlikely and even if I get it working, not practical because of minimum context. But just asking here as a last effort. I'm able to

I love Dash2Dock Lite, but Icedman is always a week or two before updates. I randomly ran pacman for the first time in weeks, not realizing GNOME had a major update. I told it to fix it in DeepSeek Ha

I have an app that processes ~500 different documents every day. An agent analyzes them, sorts them, tags them, extracts various legal nuances, creates descriptions, etc. So far I've been using Qwen 3

I'm genuinely shocked! I was able to run Qwen3.8-Flash-Next-oQ4e-mtp on M3Max 64GB with oMLX! Even a couple of weeks ago, I wasn't able to get it to run. I just tried the latest commit for fun, and it

It's been 4 years since c.ai first hallucination model, and yet, we're nowhere good enough at LLMs in terms of spontaneity/interesting hallucination features. Fortunately, last year brought new archit

arXiv:2610.13325v1 Announce Type: cross Abstract: Agentic workloads resend long contexts that barely change between calls. Prefix caching now skips most of their prefill, yet every request must still

arXiv:2610.13708v1 Announce Type: new Abstract: Long-context serving runs out of KV memory before the hardware reaches its useful throughput. HiSparse stores history in host memory and fetches only th

Basalt: Blackwell inference engine Hi all! So over the last week I've been working on a Strata fork that's heavily tuned and can achieve throughput up to 2.6x what Strata usually does on the same weig

Alright, I learend a few valuable lessons since I posted about a week ago, and I've made more strides with the DDR4x Ascend 96GB 310P cards I purchased, so here we go, meet the new friendly and less v

I stumbled across this as I was currently having glm5.3 flash run only around 10tok/s basically unusable. With Maya i am running 30 tok/s now. GLM5.3 feels even with this low quant like a much more en

DISCLAIMER: the post and the model card was made with the assist of AI. Hiya, I’m releasing Recursive BitNet N-Gram 102M, a small experiment combining ternary weights, shared transformer layers, and h

It's been 4 years since c.ai first hallucination model, and yet, we're nowhere good enough at LLMs in terms of spontaneity/interesting hallucination features. Fortunately, last year brought new archit

arXiv:2610.13325v1 Announce Type: cross Abstract: Agentic workloads resend long contexts that barely change between calls. Prefix caching now skips most of their prefill, yet every request must still

arXiv:2610.13041v1 Announce Type: new Abstract: Recurrent-depth language models improve inference-time reasoning by repeatedly applying a shared network to refine internal representations, but the add

Write-up, serve command and raw data: https://github.com/lukolszewski/vllm This shows vLLM fixes and setup that let you put PLE on SSD and save 108GB of system RAM, use int8 kv and squeeze more users