Query-Adaptive Neuron Steering for LLMs
arXiv:2610.14519v1 Announce Type: new Abstract: Effectively eliciting the capabilities of large language models (LLMs) at inference time remains an important challenge. Existing studies address this c
Knowledge catalogue
arXiv:2610.14519v1 Announce Type: new Abstract: Effectively eliciting the capabilities of large language models (LLMs) at inference time remains an important challenge. Existing studies address this c
Sampling: temp 1.0, top_p 0.95, top_k 20 for both. Run with evalscope over the OpenAI-compatible API: Qwen3.8-Next on Strata (MI50 32GB), Qwen3.8-27B on llama.cpp (RX 7900 XTX 24GB). Caveat: These are
arXiv:2605.06046v2 Announce Type: replace Abstract: Auto-regressive token generation in large language models is memory-bound because it requires 'attending to' key and value tensors (KV cache) of all
Scaling code generation without scaling review risks turning coding gains into a larger “review” backlog (whether during PR review or later, when dealing with technical debt, incidents, or bugs). The
arXiv:2610.16297v1 Announce Type: new Abstract: Scaling laws predict large-model performance from small experiments, but finding a reliable law itself requires costly experiments. SLDBench tests wheth
As strata is somewhat new, I assume that some people din't yet learn how to finetune it for best performance on their machine. I find documentation a bit hard to navigate (use your agents for this!),
arXiv:2610.14105v1 Announce Type: cross Abstract: Accurate artery extraction from X-ray coronary angiography (XCA) is essential for coronary artery disease diagnosis and intervention. However, low ves
arXiv:2610.16977v1 Announce Type: new Abstract: Visual in-context learning (ICL) enables vision-language models (VLMs) to adapt at test time via demonstration examples, but its effectiveness depends c
Today I'm excited to share that New Material is joining @MistralAI to design sovereign frontier AI. We're establishing a new team, 'Solutions Design' within Mistral: working across product, engineerin
You can't solve real-world physical tasks through autoregressive sampling alone. The physical world is entirely unforgiving to hallucination or approximate pattern matching. If we want actual autonomo
Banger paper from Google DeepMind on co-evolving an agent's harness and its weights. (bookmark it) If you already log your harness changes and results, those logs can become training data for the mode
David Silver is one of the all-time greats. He was the creator of AlphaGo, which changed the course of history by showing that neural nets could achieve superhuman performance for the first time. He a
A coding agent finishes a change and wants to open a GitHub issue with a link to its results. The agent can describe the issue, but creating it belongs to a different system. It needs a tool that can
One of the frustrating things about local AI generation on Mac is getting different models and tools to work together smoothly and fast. Valtz is a new native macOS app powered by Vpipe, a high-perfor
This post is to let people know about a workshop/tutorial on Machine Learning that will run on the (blue) math discord server: discord.gg/maths. The place is nice, people are cool and helpful, and as
If the agent isn't using my diagnostics library, I could introspect and update AGENTS.md, but there's still a delta, and it'll drift. Instead, I write a pre-commit hook that checks the git diff for di
Trying to read on what the real tested downsides have been but information seems to go both ways, some proclaim no loss, others pretty huge ones. I had some weird looping issue in their chat, and some
Suggestion for you: Pick an agent for work, and pick an agent for personal. You could explore to do this all with Grok Bot (with many different types of bots) Personally, I like @bot for work, and @Mu
I am an independent researcher. Due to budget constraints, I shifted my 21B MoE pre-training—which I had originally planned to run on 64 H100s—to a single NVIDIA GH200 ($2.29 per hour), and I have com
arXiv:2610.13361v1 Announce Type: cross Abstract: As the production of software passes into the hands of software itself, coding agents are increasingly commissioned to build agents fitted to their us
arXiv:2610.12494v1 Announce Type: new Abstract: In this article, we propose algebraic integer number formats (AINF) as a framework for highly configurable and accurate compressed computation. A key pr
arXiv:2610.13057v1 Announce Type: cross Abstract: Counterfactual text editing is the ability to revise a given text in light of a hypothetical change while preserving facts not causally affected by th
arXiv:2610.13116v1 Announce Type: new Abstract: As large language models (LLMs) acquire advanced biological design capabilities, they offer substantial research benefits but simultaneously introduce s
arXiv:2610.12507v1 Announce Type: cross Abstract: Sensing-aided beam prediction can reduce the beam-training overhead of millimeter-wave (mmWave) systems by leveraging environmental information from c
A few weeks ago we posted the open-source Graphsignal profiler here: wrap vllm serve / sglang serve / whatever in graphsignal-run and it serves everything it measures as JSON, mainly so an agent can r
arXiv:2610.13164v1 Announce Type: new Abstract: Scaling robot manipulation policies requires large amounts of diverse, action-labeled data, yet collecting such data in the real world remains expensive
arXiv:2602.02473v2 Announce Type: replace-cross Abstract: Enabling humanoid robots to perform agile and adaptive interactions has long been a core challenge in robotics. Current approaches are bottlen
I've been investigating some serious ChatGPT reliability issues over the past several days, and my latest tests revealed something particularly interesting: Both GPT-5.6 and GPT-6 showed problems unde
Overview Since introducing zero-shot time-series forecasting in BigQuery with AI.FORECAST — initially powered by TimesFM 2.0 and later upgraded to TimesFM 2.5 — you have been able to generate forecast
arXiv:2610.12738v1 Announce Type: cross Abstract: Forecasting rare, high-cost events in multivariate time-series systems is hard because their precursors are seldom observed, so a purely data-driven m
https://preview.redd.it/3ekngxk224vh1.png?width=1080&format=png&auto=webp&s=9bfe5c92033e17d5b2fb4529f1320c99455a138b I built a GPU compiler and wrote an article about it, focusing on concepts shared b
arXiv:2610.12794v1 Announce Type: cross Abstract: Large Language Model (LLM)-guided evolutionary optimization has successfully improved prompts, programs, and even entire AI systems. Most optimization
arXiv:2610.13447v1 Announce Type: cross Abstract: Speech recognition on low-cost microcontrollers is, in practice, limited to keyword spotting or a fixed list of commands. We present Oido, an inferenc
arXiv:2610.13485v1 Announce Type: new Abstract: Agentic orchestration delegates tasks to sub-agents with independent contexts, providing a key path to extending the effective execution horizon of larg
I get a lot of hate whenever I tell people I run a Q4 quant of Qwen3.8 27B with 262k KV 80t/s with vision, all while running 2 screens on windows on the same card. People seem to think I am lying or m
arXiv:2610.12532v1 Announce Type: cross Abstract: Generative creative interfaces routinely report the user's position within a space of possibilities. Such displays imply that the quantities shown wer
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
As a business intelligence user, this probably sounds familiar: Most dashboards only provide you with a standardized canvas and a fixed library of charts, and the “analysis” consists of a CSV export t
Most conversations I see on various AI subreddits start from the assumption that AGI/RSI/ASI is possible, we're on the way right now, and that things will be incredibly different afterward. The only q
indexes PDFs, Office files, scans, screenshots, audio and video, and finds them by content. A query like 'when did they decide to freeze hiring' opens the recording at that second. * EmbeddingGemma 2
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 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
'a corollary is that information about product quality will become more important and impactful' This feels like the next evolution of SEO/AEO that every product will try to game. Even if agents do th
I recently started using local LLM for mostly coding tasks for my personal projects. I also have found local models responding much better for topics such as health, fitness, ergonomics etc. compared
Standard Bots develops AI‑native industrial robot arms (used for machine tending, welding, assembly) that rely on a shared pretrained base model with 1–2 billion parameters. The model is adapted to ea
Around 3 weeks ago, I posted on this subreddit about how Jev used my one-year-old architecture, og post: https://www.reddit.com/r/LocalLLaMA/s/VpuMJt577S Then, just after that, I introduced Laya, the
Spent the past month tweaking and experimenting with many different numbers to achieve 30tps. Hardware: -Rx6700xt 12gb vram (AMD) -2x16 ddr4 3200 ram -r5 5600x -llama.cpp vulkan sdk -window11 (no wsl
Pay close attention to custom harnesses. This is a super interesting paper showing the impact of training inside the harness. Qwen3-14B goes from 22.2% to 54.8% on Spider 2.0-SQLite when it is trained
Been seeing quite a few posts about Strata lately, so I figured I'd give it a shot on my RTX PRO 6000 and I am very impressed. Most of the benchmarks I've seen are using IQ2 or IQ3 quants, so I wanted
I shipped a new feature for my blog today: the Newsletters page, which offers an index of all of the newsletters I've sent out, both my free weekly Substack and my monthly sponsors-only updates. I bui
arXiv:2610.10580v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly being integrated into hardware verification to automate specification interpretation, verification-artif
arXiv:2610.12374v1 Announce Type: new Abstract: Interactive virtual worlds allow agents to learn through exploration and interaction. What agents can learn is bounded by the environments they practice
arXiv:2610.10651v1 Announce Type: cross Abstract: This paper exposes the ergodic ceiling and thermodynamic inefficiency of current deep learning, which converges to a statistical average of historic h
arXiv:2610.12341v1 Announce Type: new Abstract: Adversarial games have driven advances from heuristic search to reinforcement learning, yet learning and adapting strategies from limited samples remain
arXiv:2610.11447v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed as coding agents that edit files, run builds and tests, inspect execution results, and repair s
arXiv:2610.11845v1 Announce Type: new Abstract: Spin in clinical trials includes reporting practices that distort the presentation of results. This is particularly critical in medicine, where spin is
arXiv:2610.11284v1 Announce Type: cross Abstract: Diffusion-based LLMs (dLLMs) have recently emerged as a promising alternative to autoregressive (AR) LLMs by enabling bidirectional parallel refinemen
arXiv:2610.12369v1 Announce Type: cross Abstract: Most robot policies keep a model in the control loop: a VLA maps observations to actions, and an Agent Harness, such as Agent-as-Policy or Harness VLA
arXiv:2610.11986v1 Announce Type: new Abstract: Vision-language models (VLMs) show great potential for damage assessment after a disaster, but a recurring deficiency is that they are reluctant to decl
Last month I created a portable memory format as a .txt file as a small project. I named it as: Memory Journal, in a shape of a skill card. It's essentially a large prompt wrapped in a text file. Init