Search: “context”

AdaSpark: Adaptive Sparsity for Efficient Long-Video Understanding

arXiv:2604.08077v1 Announce Type: new Abstract: Processing long-form videos with Video Large Language Models (Video-LLMs) is computationally prohibitive. Current efficiency methods often compromise fi

researcharxiv-cs-cv

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

Blending Human and LLM Expertise to Detect Hallucinations and Omissions in Mental Health Chatbot Responses

DGX agent

arXiv:2604.06216v1 Announce Type: cross Abstract: As LLM-powered chatbots are increasingly deployed in mental health services, detecting hallucinations and omissions has become critical for user safet

model-releasesarxiv-cs-ai

10 Apr 2026

Applications

Gaze to Insight: A Scalable AI Approach for Detecting Gaze Behaviours in Face-to-Face Collaborative Learning

arXiv:2604.03317v2 Announce Type: replace Abstract: Previous studies have illustrated the potential of analysing gaze behaviours in collaborative learning to provide educationally meaningful informati

applicationsarxiv-cs-cv

Agents

MCP is very much alive! With co-creator of MCP @dsp_ at AIE London

The referenced tweet (status ID 2042696610614849548) is not publicly accessible via search, and the X.com URL requires JavaScript/login to render. Based on the available context from related search...

agentsswyx--x

What Makes an Ideal Quote? Recommending 'Unexpected yet Rational' Quotations via Novelty

arXiv:2602.22220v2 Announce Type: replace-cross Abstract: Quotation recommendation aims to enrich writing by suggesting quotes that complement a given context, yet existing systems mostly optimize sur

safetyarxiv-cs-ai

Otaku — an LLM frontend for creative writing

I understand this kind of product may not be exactly right for this subreddit, but I still think it deserves some attention, as it's an LLM frontend for local engines in the first place, and can be us

17 Oct 2026

CARE-SD: Explainable Reasoning Architecture for Local SLM Self-Disclosure Analysis in Textual Data

arXiv:2610.21151v1 Announce Type: cross Abstract: Cybersecurity interviews uncover critical security experiences, behaviors, and privacy concerns often not fully captured by conventional technical ass

DropKV: Decoupling Residual-Output Perturbation for Near-Optimal KV-Cache Eviction

arXiv:2610.22130v1 Announce Type: new Abstract: Key-value (KV) cache management has become a critical bottleneck for scaling Large Language Models (LLMs) to long-context scenarios due to linear memory

Entertain, then Cancel: How Thinking Makes Reasoning Models Read Some and Or Literally

arXiv:2610.21204v1 Announce Type: new Abstract: When a speaker says ``some students passed'', listeners infer that not all of them did. Reasoning models now let users decide how long a model thinks be

FlashAttention with Statistics: Low-Overhead Attention Features for Compact Hallucination Detectors

arXiv:2610.21483v1 Announce Type: new Abstract: Large language models can produce fluent answers that the supplied context does not support. A compact frozen proxy model can detect such answers by rea

Functional Subnetwork Localization for Efficient Continual Learning in Large Language Models

arXiv:2610.21537v1 Announce Type: new Abstract: Biological intelligence continually adapts to new information while retaining prior skills through localized plasticity. In contrast, continual learning

On-Policy Residual Learning for Diffusion Language Models

arXiv:2610.20687v1 Announce Type: new Abstract: Residual context is key to unlocking the speed-quality Pareto frontier of diffusion language models (DLMs) by carrying forward predictive distributions

safetyarxiv-cs-lg

reflection/Beam-1.0-501B-A23B · Hugging Face

Beam is an open model built by Reflection. Beam takes text as input and generates text as output. This release includes the instruction-tuned model in BF16, FP8, and NVFP4 numerics. Beam uses a sparse

Thinking-Free Policy Initialization Makes Distilled Reasoning Models More Effective and Efficient Reasoners

arXiv:2509.26226v3 Announce Type: replace-cross Abstract: Reinforcement Learning with Verifiable Reward (RLVR) effectively solves complex tasks but demands extremely long context lengths during traini

safetyarxiv-cs-cl

What Coding Agents Leave Unsaid: Measuring Task-State Gaps Between Execution and Reports

arXiv:2610.20003v2 Announce Type: replace Abstract: Human supervisors rely on coding agents' reports to assess progress, verify repairs, and decide when to intervene. But a supervisor cannot investiga

Copy-Paste Language Models

arXiv:2610.19247v1 Announce Type: new Abstract: Transformers are intuitively considered to be efficient at copying from their inputs. We actually show that the ability of LLMs to copy in context can b

researcharxiv-cs-cl

Crossing the Gender Binary: Modelling Gendered Reference in Science Fiction and General Fiction

arXiv:2610.19028v1 Announce Type: new Abstract: Science fiction is widely credited with unsettling gender. This article asks whether that holds at the linguistic level, in how readily gendered referen

researcharxiv-cs-cl

Exposing AI Inference as a Telco Network API over the 5G User Plane

arXiv:2610.15062v1 Announce Type: new Abstract: Supporting edge Artificial Intelligence (AI) inference as a network service is challenging because current mobile systems can transport tensors but do n

How Visko streams real-time world models on Google Cloud TPUs

model-releasesgoogle-cloud-ai

Editor's note: Generative AI startups are moving rapidly beyond static clip generation to build interactive, real-time worlds. In this technical deep dive, we explore how AI startup Visko collaborated

Hardware

JetSpec++: Foresight Parallel Drafting with Self Refinements

arXiv:2610.19792v1 Announce Type: new Abstract: Speculative decoding's effectiveness is governed by two quantities: the acceptance rate of drafted tokens and the cost of producing them. We show the tw

hardwarearxiv-cs-cl

MHWA: Multi-timescale Hierarchical World-Action Model

arXiv:2610.14727v1 Announce Type: new Abstract: Autoregressive world models have emerged as a powerful paradigm for sample efficient decision making, yet their open loop rollouts often exhibit increas

Local Ai

Remember, Revisit, Revise: Scaffold Memory for Long-Horizon 3D Geometry

arXiv:2610.19674v1 Announce Type: new Abstract: Geometry foundation models estimate camera poses and 3D scene structure from video sequences, but maintaining consistent reconstructions over long horiz

local-aiarxiv-cs-cv

Hardware

Scaling LLM Inference via Parallelism: Throughput-Latency Trade-off

arXiv:2610.19405v1 Announce Type: cross Abstract: Large Language Model (LLM) serving systems must balance latency and throughput while efficiently utilizing distributed GPU resources. This challenge a

hardwarearxiv-cs-lg

SciRerankBench: Benchmarking Rerankers Towards Scientific Retrieval-Augmented Generation with LLMs

arXiv:2508.08742v3 Announce Type: replace-cross Abstract: Scientific literature question answering is a pivotal step towards new scientific discoveries. Recently, extit{two-stage} retrieval-augmented

Very impressed with LFM2.5-24-A2B in Pi Agent

Just downloaded the new LFm2.5-24B-A2B LiquidAI model for testing. They claimed some pretty good performance, and I've had good experiences with their prior releases. Some context - I use Ornith 1.5-3

What makes NVIDIA Nemotron 3.5 Super VL different? → Text, image, document, and video understanding in one model → Reasoning and tool use wi…

model-releasestogether-ai--x

What makes NVIDIA Nemotron 3.5 Super VL different? → Text, image, document, and video understanding in one model → Reasoning and tool use with visual context preserved across steps → 120B hybrid MoE (

Where Do LLM Ensemble Gains Come From? Decomposing Candidate Value, Diversity, and Aggregation

arXiv:2610.18786v1 Announce Type: cross Abstract: LLM ensemble gains are often measured against a single baseline, but the value attributed to additional candidates can change with the comparison used

Foundation Models for Event Prediction Across Heterogeneous Event Logs

model-releasesarxiv-cs-lg

arXiv:2610.17503v1 Announce Type: new Abstract: Event logs are ubiquitous across real-world domains, including business processes, manufacturing workflows, healthcare, and many other operational syste

I built a cheap dual 3090 rig from used parts, but why isn't it much faster?

Hey everyone, I recently put together a local LLM machine using mostly second hand hardware. I wanted to experiment with local models for app development without spending a fortune on a new workstatio

Tutorials

I trained a 15M 'decision-only' model two ways — linear recurrence vs Transformer attention. Accuracy was a coin flip; latency and training cost weren't.[R]

tutorialsr-machinelearning

Inspired by TypeSafe's Jev (a non-generative model that only outputs bounded decisions — choices, probabilities, calibrated confidence), I built two small decision models end-to-end on identical budge

InfraWorld: Motion-Aware World Modeling for Roadside 4D Occupancy Forecasting

model-releasesarxiv-cs-cv

arXiv:2610.14544v1 Announce Type: new Abstract: Roadside 4D occupancy forecasting predicts future semantic occupancy in a fixed infrastructure coordinate system. Roadside scenes are globally structure

MultiAlloc: Budget-Aware Precision Allocation for KV Cache Quantization

model-releasesarxiv-cs-lg

arXiv:2610.18642v1 Announce Type: new Abstract: The Key-Value (KV) cache constitutes a dominant source of memory consumption during large language model (LLM) inference, particularly in long-context a

Qwen 3.8 Flash Next, DeepSeekV4 Flash and Time

MacStudio 128GB M1 ultra I'm curious how people who can run both models are deciding which one to work with. DeepSeekv4 benchmarks right below GPT 5.5 high. Flash Next benchmarks right above 5.4/5.5 x

RIFT: Token Influence Follows the Path of Least Resistance

model-releasesarxiv-cs-lg

arXiv:2610.16349v1 Announce Type: cross Abstract: Which context tokens drive a language model's answer? Graph-based tools such as attention rollout and attention flow answer this from a single forward

Which Trajectory Is Safer? Combining Rule and Learning for Safety Evaluation

arXiv:2610.17806v1 Announce Type: new Abstract: Collision rate remains a dominant metric for evaluating autonomous driving trajectory safety, but its binary nature obscures substantial safety differen

safetyarxiv-cs-ro

Horsey challenge | Strata Fork | Qwen 3.8 FN NVFP4

So I wanted to try the horsey challenge myself on the latest engine i been testing which is a strata fork by sergqwer. https://github.com/sergqwer/strata-nvfp4. I seen a post earlier and seen i believ

13 Oct 2026

GIST: Generator-state-Informed Supervision for Tabular Foundation Models

arXiv:2610.12904v1 Announce Type: new Abstract: Tabular foundation models (TFMs) based on prior-data fitted networks typically use sampled hard query labels, leaving avoidable sampling variability in

researcharxiv-cs-lg

KV-Repurpose: A Testbed for Evaluating Non-Prefix KV-Cache Reuse

arXiv:2610.13060v1 Announce Type: new Abstract: KV-cache reuse in large language model serving is commonly limited to exact prompt prefixes, preventing reuse when the same retrieved document appears a

researcharxiv-cs-cl

Local Ai

LLC: lightweight Open WebUI alt now with Agents+MCP+RAG+Image Editing

Quick context: LocalLightChat (LLC) is a chat frontend for local LLMs. Download it, run it, that's it. No installation (unless you want to), no dependencies, and it runs on pretty much anything, inclu

local-air-localllama

strix-llama new branch: prompt processing +21–54% over Vulkan and up to 2.7× over stock ROCm kernel at 128k, same outputs quality

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

Building a 4x R9700 setup for a 10 person startup

Just wanted to share a build I am doing for a client. $18k. Specs: Threadripper 9970x 128gb DDR5 ECC 5600 4x AMD Radeon R9700 AI Pro 32gb(128gb total VRAM) 1600W PSU(GPUs are undervolted and under 210

11 Oct 2026

UPDATE: Qwen 3.8 27B 140 tok/s on single RTX 3090 Megakernel: KL divergence 0.0009 vs llama.cpp

This is a follow-up to my post from yesterday (https://www.reddit.com/r/LocalLLaMA/comments/1x2erdj/qwen3827b_on_a_single_3090_140_toks_on_code_with/) detailing the CUDA megakernel. Thank you all for

11 Oct 2026

Improve token per second without touching quant

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

10 Oct 2026

BioBigBird: A Sparse Attention Model for Long-Range Dependency Processing in Biomedical Text

model-releasesarxiv-cs-cl

arXiv:2610.11430v1 Announce Type: new Abstract: While domain-specific Large Language Models (LLMs) have encoded vast biomedical knowledge, their limited context windows often hinder a deep understandi

Look Back, Think Ahead: Visual Memory on Demand for Efficient Multimodal Reasoning

arXiv:2610.12060v1 Announce Type: new Abstract: Processing long visual token sequences from high-resolution images makes multi-step reasoning computationally expensive for multimodal Large Language Mo

researcharxiv-cs-cv

Agents

MemTrace: State-Consistent Memory for Long-Horizon Coding Agents

arXiv:2610.04838v2 Announce Type: replace Abstract: As coding agents take on long-horizon software evolution tasks spanning multiple files and stages, longer execution trajectories introduce two coupl

agentsarxiv-cs-ai

RaReCache: Bridging the Gap in Cross-Model KV Cache Reuse via Rank disagreement-based Selective Recomputation

arXiv:2610.11358v1 Announce Type: new Abstract: Cross-model KV-cache reuse remains a key challenge in modern LLM serving. Coding agents and multi-model systems increasingly route a shared context acro

Real Long-Term Memory for AI: A 50-Million-Token Window That Is Faster and Cheaper Than Recompute

arXiv:2610.10845v1 Announce Type: cross Abstract: A large language model can only use the text that fits in its context window, and it recomputes its internal key-value (KV) state for a prompt every t