Knowledge catalogue
arXiv:2604.16368v1 Announce Type: new Abstract: Speculative decoding accelerates LLM inference by using a small draft model to propose k candidate tokens for a target model to verify. While effective
arXiv:2604.16380v1 Announce Type: new Abstract: Large language models (LLMs) rely on pretraining on massive and heterogeneous corpora, where training data composition has a decisive impact on training
arXiv:2601.10306v2 Announce Type: replace-cross Abstract: While Reinforcement Learning (RL) has advanced LLM reasoning, applying it to long-context scenarios is hindered by sparsity of outcome rewards
arXiv:2604.16778v1 Announce Type: new Abstract: LLM-powered agents often reason from scratch when presented with a new problem instance and lack automatic mechanisms to transfer learned skills to othe
Today, we're excited to announce Claude Cowork in Amazon Bedrock. You can now run Cowork and Claude Code Desktop through Amazon Bedrock, directly or using an LLM gateway. In this post, we walk through
arXiv:2601.21278v2 Announce Type: replace-cross Abstract: Vision Language Models (VLMs) are good at recognizing the global location of a photograph -- their geolocation prediction accuracy rivals the
arXiv:2604.16377v1 Announce Type: new Abstract: Large Language Models (LLMs) trained on massive code corpora are now increasingly capable of generating code that is hard to distinguish from human-writ
arXiv:2510.06700v3 Announce Type: replace Abstract: Both humans and large language models (LLMs) exhibit content effects: biases in which the plausibility of the semantic content of a reasoning proble
This article explains how to ground Korean AI agents in realistic demographic data by utilizing synthetic personas, likely leveraging NVIDIA's Nemotron models available on Hugging Face. The approach e
arXiv:2603.16120v2 Announce Type: replace Abstract: Deep Research (DR) systems help researchers cope with ballooning publishing counts. Such tools synthesize scientific papers to answer research queri
arXiv:2604.17897v1 Announce Type: new Abstract: The advent of Large Language Models (LLMs) has fundamentally reshaped the way we interact with graphs, giving rise to a new paradigm called GraphLLM. As
arXiv:2511.21686v2 Announce Type: replace Abstract: Synthetic data has become increasingly important for training large language models, especially when real data is scarce, expensive, or privacy-sens
arXiv:2510.09474v2 Announce Type: replace Abstract: Modern conversational agents like ChatGPT and Alexa+ rely on predefined policies specifying metadata, response styles, and tool-usage rules. As thes
arXiv:2505.17238v3 Announce Type: replace Abstract: Collaborative dialogue offers rich insights into students' learning and critical thinking, which is essential for personalizing pedagogical agent in
arXiv:2604.17132v1 Announce Type: new Abstract: Safety-aligned large language models (LLMs) often generate refusal responses to harmless queries due to the over-refusal problem. However, existing meth
arXiv:2604.17837v1 Announce Type: cross Abstract: An LLM's residual stream is both state and instruction: it encodes the current context and determines the next transformation. We introduce a paramete
arXiv:2604.17175v1 Announce Type: new Abstract: We introduce RosettaSearch, an inference-time multi-objective optimization approach for protein sequence optimization. We use large language models (LLM
arXiv:2604.16606v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed in high-stakes domains, yet a unified treatment of their overlapping safety challenges remains
arXiv:2604.16423v1 Announce Type: new Abstract: Defensive training methods such as positive preventative steering (PPS) and inoculation prompting (IP) offer surprising results through seemingly simila
arXiv:2604.04804v2 Announce Type: replace Abstract: Learning from experience is critical for building capable large language model (LLM) agents, yet prevailing self-evolving paradigms remain inefficie
arXiv:2509.24328v2 Announce Type: replace Abstract: LLMs have low GPU efficiency and high latency due to autoregressive decoding. Speculative decoding (SD) mitigates this using a small draft model to
arXiv:2604.18401v1 Announce Type: new Abstract: General agents have given rise to phenomenal applications such as OpenClaw and Claude Code. As these agent systems (a.k.a. Harnesses) strive for bolder
arXiv:2604.17159v1 Announce Type: cross Abstract: We present, to our knowledge, the most comprehensive cross-model evaluation of LLM agents on offensive cybersecurity tasks, benchmarking 10 frontier m
arXiv:2504.18269v2 Announce Type: replace Abstract: When generating images from prompts that include specific entities, the model must retain as much entity-specific knowledge as possible. However, th
arXiv:2604.18096v1 Announce Type: cross Abstract: LLMs are increasingly presented as collaborators in programming, design, writing, and analysis. Yet the practical experience of working with them ofte
arXiv:2604.18493v1 Announce Type: new Abstract: Reinforcement Learning (RL) enhances LLM reasoning, yet a paradox emerges as models scale: strong base models saturate standard benchmarks (e.g., MATH),
arXiv:2604.18539v1 Announce Type: new Abstract: This paper studies how empirical dialogue-flow statistics can be incorporated into Next Dialogue Act Prediction (NDAP). A KL regularization term is prop
We're open-sourcing FlashKDA — our high-performance CUTLASS-based implementation of Kimi Delta Attention kernels. Achieves 1.72×–2.22× prefill speedup over the flash-linear-attention baseline on H20,
arXiv:2510.17795v3 Announce Type: replace Abstract: Replicating AI research is a crucial yet challenging task for large language model (LLM) agents. Existing approaches often struggle to generate exec
arXiv:2604.17422v1 Announce Type: new Abstract: Long video understanding remains a formidable challenge for Multimodal Large Language Models (MLLMs) due to the prohibitive computational cost of proces
OpenAI released ChatGPT Images 2.0 today, their latest image generation model. On the livestream Sam Altman said that the leap from gpt-image-1 to gpt-image-2 was equivalent to jumping from GPT-3 to G
arXiv:2604.16286v1 Announce Type: new Abstract: As AI systems are increasingly used to conduct research autonomously, misaligned systems could introduce subtle flaws that produce misleading results wh
B8852 is a release of llama.cpp, an LLM inference library in C/C++ . This build tag likely contains bug fixes, performance improvements, and feature updates from the llama.cpp development cycle, which
b8853 is a release of llama.cpp , the open-source C/C++ library for running large language model inference. The project enables LLM inference with minimal setup and state-of-the-art performance on a w
B8857 is an intermediate build release from the llama.cpp project, which is a C/C++ implementation that enables LLM inference with minimal setup and state-of-the-art performance on a wide range of har
arXiv:2604.15503v1 Announce Type: new Abstract: Recent breakthroughs in language models (LMs) using neural networks have raised the question: how similar are these models' processing to human language
arXiv:2503.16505v4 Announce Type: replace-cross Abstract: A critical challenge in social science research is the high cost associated with experiments involving human participants. We identify Synthet
arXiv:2512.17052v4 Announce Type: replace Abstract: Function calling agents powered by Large Language Models (LLMs) select external tools to automate complex tasks. On-device agents typically use a re
arXiv:2602.01566v2 Announce Type: replace Abstract: Deep research is emerging as a representative long-horizon task for large language model (LLM) agents. However, long trajectories in deep research o
arXiv:2512.12858v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are increasingly deployed in business-critical domains such as finance, education, healthcare, and customer suppo
arXiv:2604.15760v1 Announce Type: new Abstract: We introduce the first version of KWBench (Knowledge Work Bench), a benchmark for unprompted problem recognition in large language models: can an LLM id
arXiv:2604.15385v1 Announce Type: cross Abstract: Software documentation is essential for program comprehension, developer onboarding, code review, and long-term maintenance. Yet producing quality doc
arXiv:2604.16263v1 Announce Type: new Abstract: Coordinating multi-robot systems (MRS) to search in unknown environments is particularly challenging for tasks that require semantic reasoning beyond ge
arXiv:2603.19339v2 Announce Type: replace-cross Abstract: Dimensionality reduction is critical for deploying dense retrieval systems at scale, yet mainstream post-hoc methods face a fundamental trade-
arXiv:2604.15416v1 Announce Type: cross Abstract: Sign-based optimization algorithms, such as SignSGD, have garnered significant attention for their remarkable performance in distributed learning and
arXiv:2604.14646v2 Announce Type: replace Abstract: Recent advances in reinforcement learning (RL) have improved the reasoning capabilities of large language models (LLMs) and vision-language models (
arXiv:2604.15333v1 Announce Type: cross Abstract: Human-AI romantic relationships are increasingly common, yet little is understood about how public discourse around them emerges and shifts over time.
arXiv:2604.16280v1 Announce Type: new Abstract: Explaining Machine Learning (ML) results in a transparent and user-friendly manner remains a challenging task of Explainable Artificial Intelligence (XA
arXiv:2604.15375v1 Announce Type: cross Abstract: Large Language Models (LLMs) have shown significant improvement in RTL code generation. Despite the advances, the generated code is often riddled with
arXiv:2510.27617v2 Announce Type: replace Abstract: Automation of Register Transfer Level (RTL) design can help developers meet increasing computational demands. Large Language Models (LLMs) show prom
arXiv:2601.12193v3 Announce Type: replace Abstract: Modern video retrieval systems are expected to handle diverse tasks ranging from corpus-level retrieval, fine-grained moment localization to flexibl
arXiv:2604.15613v1 Announce Type: cross Abstract: We present VoodooNet, a non-iterative neural architecture that replaces the stochastic gradient descent (SGD) paradigm with a closed-form analytic sol
We are excited to have @baseten as a day 0 launch partner for Kimi K2.6! Their inference stack brings KV-aware routing, NVFP4 on Blackwell, multi-modal hierarchical caching, and prefill-decode disaggr
arXiv:2510.09689v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have been augmented with web search to overcome the limitations of the static knowledge boundary by accessing up-
b8841 is a release of llama.cpp dated April 18, 2026 , a C/C++ library for large language model inference. The main goal of llama.cpp is to enable LLM inference with minimal setup and state-of-the-art
b8842 is a release of llama.cpp, a C/C++ implementation for LLM inference. The llama.cpp project publishes multiple releases in a single day as part of its active development cycle. This specific rele
b8848 is a release from llama.cpp, an open-source C/C++ project for LLM inference . This release builds on the rapidly-developing codebase with regular updates that include bug fixes, feature improvem
The llama.cpp project is the main playground for developing new features for the ggml library , and the main goal of llama.cpp is to enable LLM inference with minimal setup and state-of-the-art perfor
Human cognitive friction has long been acting as a regularizer for a lot of digital infrastructure. It made software APIs less terrible and codebases less complex. Now LLM disintermediation is causing
Release b8838 is a version of llama.cpp, a C/C++ implementation for LLM inference. The release likely contains bug fixes, performance improvements, and updates to the codebase for running large langua