Blast Radius
arXiv:2608.07440v1 Announce Type: new Abstract: Agentic coding faces growing problems of affordability and wasted tokens. We introduce Blast Radius, a predictive memory management layer that estimates
Knowledge catalogue
arXiv:2608.07440v1 Announce Type: new Abstract: Agentic coding faces growing problems of affordability and wasted tokens. We introduce Blast Radius, a predictive memory management layer that estimates
Hey everyone, I've been experimenting with the new DeepSeek-V4-Flash-0731 release locally using the Unsloth Studio Q8_K_XL GGUF with OpenCode. Overall, it's been working really well, but I've noticed
arXiv:2608.05240v1 Announce Type: cross Abstract: One-bit post-training quantization represents each weight using only its sign, requiring all deployment contexts to share the same binary weight matri
arXiv:2506.13502v3 Announce Type: replace Abstract: Next-word prediction (NWP) trains language models against a single observed continuation, even though many contexts admit multiple plausible next wo
arXiv:2608.01651v1 Announce Type: cross Abstract: Hybrid-attention large language models combine full attention with recurrent linear attention to reduce long-context inference costs, yet their autore
arXiv:2607.29591v1 Announce Type: new Abstract: KV cache compression is essential for efficient long-context inference. Existing eviction methods permanently discard unselected tokens and consequently
arXiv:2607.27167v1 Announce Type: cross Abstract: LLM-based agents excel at software engineering tasks where an existing codebase provides context, but constructing a program from scratch remains fund
arXiv:2607.23634v1 Announce Type: cross Abstract: Attention enables context modeling via query-key scoring with softmax normalization. Driven by industrial long-context demands, mainstream research ha
arXiv:2607.20368v1 Announce Type: new Abstract: Recent autoregressive video diffusion methods are increasingly built upon Self Forcing, where the student is trained on histories produced by its own ro
arXiv:2607.12161v1 Announce Type: new Abstract: Context-reduction layers for API-based coding agents, including command-output compressors, retrieval rankers, and payload-optimizing proxies, are usual
Editor’s note: Some of the most interesting questions in AI are being asked by information theoreticians, around how to provide context to an emerging class of AI agents. A few weeks ago, we waded int
arXiv:2607.02800v1 Announce Type: new Abstract: Induction heads are attention circuits believed to underlie in-context learning in transformers, yet a precise characterization of the estimators they i
arXiv:2603.02112v2 Announce Type: replace-cross Abstract: Modern language models reason within bounded context, an inherent constraint that poses a fundamental barrier to long-horizon reasoning. We id
In the agentic era, the role of the database has fundamentally changed. It is no longer a passive repository; it’s a critical context engine designed to ground generative AI apps, models and power aut
arXiv:2606.11173v1 Announce Type: new Abstract: Conditioning a language model on additional context, such as feedback on a previous attempt, typically improves its response. Self-distillation trains t
arXiv:2606.06467v1 Announce Type: new Abstract: Long-context inference in modern LLMs is increasingly constrained by decoding efficiency, especially in reasoning-heavy settings where models generate l
arXiv:2606.00408v1 Announce Type: cross Abstract: Long-horizon search agents accumulate large amounts of retrieved content across many tool calls, making context-budget efficiency increasingly importa
arXiv:2602.16284v2 Announce Type: replace Abstract: Scaling language models to long contexts is often bottlenecked by the size of the key-value (KV) cache. In deployed settings, long contexts are typi
arXiv:2605.16113v1 Announce Type: cross Abstract: Large language models (LLMs) have achieved unprecedented success due to their exceptional generative capabilities. However, because they depend on kno
arXiv:2605.12260v1 Announce Type: new Abstract: Long-horizon language agents accumulate conversation history far faster than any fixed context window can hold, making memory management critical to bot
To act at the speed of business, AI agents must operate in fast and trusted reasoning loops. They need to “think” by reasoning across both your historical context and your live operational reality. On
arXiv:2604.19678v1 Announce Type: new Abstract: Function vectors (FVs) are vector representations of tasks extracted from model activations during in-context learning. While prior work has shown that
This is awesome and makes things much more composable. Love it deepagents now supports structured output for subagents! an under appreciated piece of context engineering is figuring out what context i
arXiv:2604.08782v1 Announce Type: new Abstract: Large language models (LLMs) suffer significant performance degradation when user instructions and context are distributed over multiple conversational
arXiv:2509.24169v2 Announce Type: replace Abstract: Large Language Models (LLMs) can perform new tasks from in-context demonstrations, a phenomenon known as in-context learning (ICL). Recent work sugg
In production systems today, most AI agents operate as lightweight dispatchers. They parse a prompt, make an API call, retrieve a record, and hand back control. Their execution is fast, stateless and
Today, Reflection released Beam, its first open-weight model built to give developers, researchers, and enterprises open control and customization for frontier software engineering and long-horizon ag
arXiv:2610.19709v1 Announce Type: new Abstract: Mamba models underperform Transformers on in-context retrieval. This gap is usually attributed to Mamba's fixed-size recurrent state, which loses inform
arXiv:2610.17959v1 Announce Type: cross Abstract: The design of attention mechanisms in long-context large language models faces an ``impossible triangle'' among KV cache size, training and prefill FL
arXiv:2610.15883v1 Announce Type: cross Abstract: Text wrapped in a trigger can be made behaviorally invisible to a language model: omitted from summaries and quotations, ignored in question answering
arXiv:2610.17901v1 Announce Type: new Abstract: Long-context training of large language and multimodal models couples structured attention masks (document, prefix-LM, share-question, sliding-window, b
graph: skip the LM head for pooled embeddings (#30445) graph: skip the LM head for pooled embeddings A context that extracts pooled embeddings outputs one vector per sequence, but every model still ex
arXiv:2610.08722v1 Announce Type: new Abstract: Many long-horizon agents compact their context on a global rule, usually a token budget, blind to what the agent was doing. We ask whether the agent's r
arXiv:2610.08527v1 Announce Type: new Abstract: Hybrid architectures combining linear sequence models with softmax attention provide an effective balance between efficient long-context modeling and pr
arXiv:2610.02488v1 Announce Type: new Abstract: Language-model agents increasingly rely on harnesses that manage bounded context, persistent memory, tools, verification, and repeated execution, yet ex
arXiv:2610.03135v1 Announce Type: new Abstract: Serving long-context autoregressive language models is constrained by the key-value (KV) cache. Most dynamic eviction methods score token importance per
arXiv:2610.00675v1 Announce Type: cross Abstract: Evolutionary approaches to LLM-driven discovery often generate new programs from a small set of selected ancestors. This keeps contexts manageable but
arXiv:2610.00041v1 Announce Type: cross Abstract: Agent frameworks increasingly delegate work by forking sub-agents; a common default makes the child inherit the parent's full working context. We meas
This is 6 bc-250 ex mining boards with 5 in the asrock 4u12g case they came in. After a lot of testing my current preferred setup is 4 boards running Qwen Next Flash IQ2_XS at 100k context with around
arXiv:2609.39740v1 Announce Type: new Abstract: Long-context reasoning faces two complementary bottlenecks: retaining evidence across long inputs and sustaining computation across many reasoning steps
arXiv:2609.39068v1 Announce Type: new Abstract: Long-context LLM agents accumulate interaction histories that strain KV-cache memory and attention computation. Although sparse attention reduces these
arXiv:2609.36835v1 Announce Type: new Abstract: Long-context large language model inference is bottlenecked by KV caches that grow linearly with sequence length. This burden is especially severe for l
arXiv:2609.36966v1 Announce Type: cross Abstract: Covariate effects vary across contexts and shift over time, requiring forecasters to assess how to use them for each forecasting context. As forecasti
arXiv:2609.32610v1 Announce Type: cross Abstract: Large language models represent context with a key-value (KV) cache. Caches are model-specific: for the same text, models with different architectures
arXiv:2609.31306v1 Announce Type: new Abstract: Tabular in-context learners such as TabPFN, Mitra, or ConTextTab rely on alternating row and column attention over 2D sequences of latent embeddings. Th
arXiv:2609.30854v1 Announce Type: cross Abstract: Autoregressive LLM inference at long context is bounded by memory bandwidth, not arithmetic throughput, and the binding resource shifts from model wei
Hey everyone! I want to share the results of a small language model architecture experiment I’ve been running as a side project. The ultimate goal is to test an approach to reduce the number of the mo
Welcome to the first Cloud CISO Perspectives for September 2026. Today, Sandra Joyce shares the latest details on Google’s visibility into how attackers are using AI, and how we’re using AI to stop th
arXiv:2609.17109v1 Announce Type: new Abstract: A common small-model deployment runs one shared backbone with several LoRA specialists that answer over the same context. Serving them naively re-prefil
arXiv:2607.09493v2 Announce Type: replace Abstract: Agentic LLM systems that generate code through multi-turn tool use face a fundamental context problem: each session starts from zero, discarding the
arXiv:2609.14245v1 Announce Type: cross Abstract: Context compression reduces generator input in retrieval-augmented generation, but answer quality alone does not characterize citation attribution. We
I find this new model at HF: 'Built for demanding work. A 262 144-token context window, adjustable reasoning effort, tool calling, and text, image and video understanding. Architecture Agnes-3.0-Flash
I am just sharing my config for Qwen 3.8 27b that fits on a 5060TI, what is cool about this is that you can even get vision! and a 85K context (I have 1.5gb of headroom for more context or a better qu
I like this framing a lot: agent improvement is harness improvement, not model improvement The interesting interventions are often at the tool boundary: - what context you pass - when tools are provid
arXiv:2609.03880v1 Announce Type: new Abstract: We introduce Xiaomi-TabLDM, a tabular large data foundation model for classification and regression via in-context learning, which delivers superior pre
mtmd : add const in various places (#28307) mtmd : mark context as const in more methods Mark mtmd_context as const in: mtmd_bitmap_init_lazy mtmd_tokenize mtmd_tokenize_from_parts mtmd_helper_support
arXiv:2609.01601v1 Announce Type: cross Abstract: The repository-level code generation task requires synthesizing code that satisfies task requirements while remaining consistent with the target repos
arXiv:2609.00891v1 Announce Type: new Abstract: Sharing context between LLMs in a multi-model system requires the receiving model to prefill the shared prefix because KV caches are model-specific. Rec
arXiv:2608.29571v1 Announce Type: new Abstract: Flexible adaptation to context and shared pragmatic intuitions contribute to smooth human conversation. Iterated reference games---in which players repe
ok bit more context: it's actually a QAT Q2 for Qwen 3.8 27 B: https://huggingface.co/sdkyuan/qwen3.8-27B-qat-q2_0-gguf QAT Q2 for DFlash model: https://huggingface.co/HermiHg/Qwen3.8-27B-DFlash2-Q2_K