Search: “llm”

Towards Understanding On-Policy Distillation through the Lens of Test-Time Scaling

arXiv:2608.11829v1 Announce Type: cross Abstract: On-policy distillation (OPD) has emerged as a promising post-training technique for enhancing LLM reasoning. It is commonly believed to enable the stu

safetyarxiv-cs-cl

13 Aug 2026

Human

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Tutorials

We just pushed a big update to Sakana Chat. No login required and free to use: https://chat.sakana.ai/ It is now powered by Fugu and our upd…

DGX agent

We just pushed a big update to Sakana Chat. No login required and free to use: https://chat.sakana.ai/ It is now powered by Fugu and our updated Namazu Japanese LLM, and comes with full code execution

tutorialsdavid-ha--x

13 Aug 2026

Why AI Detection Fails for Academic Integrity

arXiv:2608.11256v1 Announce Type: new Abstract: Institutions use commercial AI detectors for academic integrity, yet detectors cannot distinguish AI editing from full LLM drafts and may treat both as

safetyarxiv-cs-lg

13 Aug 2026

A Gateway Architecture for Enterprise MCP Authentication: Unifying Heterogeneous Auth, Identity Delegation, and the User / Non-User Persona Problem

arXiv:2608.10760v1 Announce Type: cross Abstract: The Model Context Protocol (MCP) has become the de-facto interface for connecting LLM agents to enterprise tools, and adoption has been explosive: wit

Beyond a Bag of Features: Set-Level Instability in Sparse Autoencoders

arXiv:2608.11197v1 Announce Type: cross Abstract: Shani et al. (2026) show that LLM representations broadly recover human category boundaries, while failing to reflect fine-grained typicality structur

Local Ai

Conversational Orchestration for Organic 6G

arXiv:2608.10714v1 Announce Type: cross Abstract: The Organic 6G vision of a network of networks spanning an edge-cloud continuum complemented by non-terrestrial resources requires, to realize its pro

local-aiarxiv-cs-ai

Cost-Efficient Estimation of General Abilities Across Benchmarks

arXiv:2604.01418v2 Announce Type: replace Abstract: Thousands of diverse benchmarks have been developed to measure the quality of large language models (LLMs). Yet prior work has demonstrated that LLM

ELMER: Evolutionary Language Model that Explores and Refines

arXiv:2608.10196v1 Announce Type: cross Abstract: Program evolution can measure whether a mutation helped, but it rarely controls how far the mutation moves in behavior space. Syntactic edit size is a

From Faulty Memories to Corrected Actions: Dependency-Guided Rollback Repair for Memory-Augmented Agents

arXiv:2608.10502v1 Announce Type: new Abstract: Persistent memory lets language-model agents reuse information across sessions, but it also makes errors durable: a poisoned, stale, or misattributed re

MAP-Graph: Provenance-Aware Shared Memory for Multi-Agent Workflows

arXiv:2608.10509v1 Announce Type: new Abstract: Shared memory helps language-model agents reuse information across long workflows, yet relevant evidence may not be admissible for a particular agent or

MarkNull: Model-Agnostic Watermark Removal in AI-Generated Images via On-Manifold Latent Manipulation

arXiv:2608.10166v1 Announce Type: cross Abstract: Digital watermarking has emerged as a critical technique for provenance and copyright attribution in AI-generated imagery, yet its robustness against

MERA: Model Evolution and Routing with Skill Adaptation for Agentic Systems at Scale

arXiv:2608.10333v1 Announce Type: new Abstract: LLM agents execute heterogeneous sequences of model calls within a single task: some invocations require careful reasoning, while others are structured

safetyarxiv-cs-lg

Mitigating Bus Bunching with Reinforcement Learning Enhanced by Semantic Stop Embedding

arXiv:2608.10207v1 Announce Type: new Abstract: Bus bunching degrades service regularity and increases passenger waiting in high-frequency transit. Existing reinforcement-learning-based holding contro

Multimodal Item Parameter Estimation using Simulated Response Probabilitie

arXiv:2608.10154v1 Announce Type: cross Abstract: We present results from reconstructing multiple-choice model (MCM) and three-parameter logistic (3PL) model curves using a fine-tuned multimodal large

MUSE: A Full-Text Cross-Domain Knowledge Base of Scientific Problems, Solutions, and Rationales

arXiv:2608.10974v1 Announce Type: new Abstract: Scientific papers contain fine-grained records of problem solving: authors mention technical obstacles and methods that were used to address them, often

Patients With Personality: Realistic Patient Simulation through Controlled Diversity and Selective Disclosure

arXiv:2606.17441v2 Announce Type: replace-cross Abstract: Simulating realistic patient interactions is a key requirement to testing clinical applications of LLMs at scale without time-consuming and ex

researcharxiv-cs-ai

PERCEPT: A Corpus for POS Tagging and Analysis of Persian-English Code-Mixing

arXiv:2608.10109v1 Announce Type: new Abstract: Social media has become a major venue for multilingual communication, where users frequently mix multiple languages within a single utterance. Although

Power law graph attention: exact generalization of scaled dot-product attention, empirical collapse at inference

arXiv:2608.10288v1 Announce Type: cross Abstract: The Large Language Model from Power Law Decoder Representations (PLDR-LLM) and its attention, Power Law Graph Attention (PLGA), replace the fixed bili

SBCO: Self-Supervised, Verifier-Grounded Harness Optimization For Planning Agents

arXiv:2608.10157v1 Announce Type: new Abstract: Self-improving agents seek to reduce the human engineering effort behind AI systems by enabling them to evolve and self-improve their performance over t

The Truth Stays in the Family: Enhancing Contextual Grounding via Inherited Truthful Heads in Model Lineages

arXiv:2606.15821v2 Announce Type: replace-cross Abstract: Recent advances in large language models (LLMs) have produced many specialized multimodal LLMs (MLLMs) that share common foundational LLMs, fo

UniMod: Enhancing Multi-Modal Medical Diagnosis through Cross-Modality and Within-Modality Alignment

arXiv:2608.10316v1 Announce Type: new Abstract: Multi-modal learning combining medical images and clinical text is promising for disease diagnosis. However, standard multi-modal training leads to shor

safetyarxiv-cs-cv

VibeLifeBench: Can Your Life Agent Be Proactive and Persistent in a Living World?

arXiv:2608.10875v1 Announce Type: cross Abstract: Large language model (LLM) agents are increasingly deployed as personal assistants. Existing evaluations, however, mostly use short, self-contained re

Local Ai

Agentic AI-driven Immersive Simulation: A Knowledge-Aware Virtual Training Platform forHigh Dose Rate (HDR) Brachytherapy

arXiv:2608.08163v1 Announce Type: new Abstract: The convergence of the Metaverse and Large Language Model (LLM)-based AI agent is catalyzing a shift toward autonomous, immersive, and personalized peda

local-aiarxiv-cs-ai

b10361

model-releasesllama-cpp-releases

model : fix SWA not being enabled for EXAONE 4.5 (#26848) model : fix SWA not being enabled for EXAONE 4.5 load_arch_hparams tests hparams.n_layer() == 64 before LLM_KV_NEXTN_PREDICT_LAYERS has been r

Tutorials

Beyond the Node: Clade-level Selection for Efficient MCTS in Automatic Heuristic Design

arXiv:2602.00549v2 Announce Type: replace Abstract: While Monte Carlo Tree Search (MCTS) shows promise in Large Language Model (LLM) based Automatic Heuristic Design (AHD), it suffers from a critical

tutorialsarxiv-cs-lg

Business Truth, not SQL Accuracy: A Rule-Gated 7B Analytics Agent Outperforms a Direct-Prompted 32B Baseline

arXiv:2608.09254v1 Announce Type: new Abstract: LLM analytics agents are evaluated on SQL syntax accuracy, but production failures look different: questions with two valid business definitions, questi

Can Gemma and Qwen models catch hallucinations by looking at their own logprobs?

model-releasesr-localllama

Hi! I'm really obsessed with LLM hallucinations for the last 6 days 😭 I started by designing system prompts to attack hallucinations but failed, obviously. Now I tried reading logprobs and... I think

DarwinX: Evolving Agent Harnesses Through Natural Selection

arXiv:2608.07545v1 Announce Type: cross Abstract: An LLM agent's capability depends not only on model weights but on its harness: prompts, tools, skills, and control flow. Self-improvement loops alrea

Different Feedback, Different Updates: Selective Self-Learning from User Interactions for Large Language Models

arXiv:2608.09109v1 Announce Type: new Abstract: User feedback offers natural supervision for persistent LLM improvement, but a single message may support multiple behavioral changes with different sco

researcharxiv-cs-ai

ForestBench: A Unified Graph Framework for Evaluating Multi-Agent Collaboration

arXiv:2608.08605v1 Announce Type: new Abstract: Multi-agent systems (MAS) built on Large Language Models (LLMs) are proliferating rapidly, but their heterogeneous execution traces provide no common ba

Human-Level Text-to-SQL via Reinforcement Learning on Verified Data, Without Pipeline Engineering

arXiv:2603.20004v3 Announce Type: replace-cross Abstract: Translating natural language questions to SQL queries (Text-to-SQL) is a long-standing problem in database research. Recent efforts have focus

Learning Preference Adaptation for Large Language Model Personalization via Verbal Reinforcement Learning

arXiv:2608.09507v1 Announce Type: cross Abstract: Natural language user preferences provide an interpretable interface for LLM personalization. However, universal preference summaries often contain in

Mismatch Matters: On-Policy Distillation Beyond Token Agreement

arXiv:2608.09836v1 Announce Type: new Abstract: On-policy distillation (OPD) has emerged as a core component of modern LLM post-training pipelines, yet we reveal a failure mode: degenerate agreement,

Mitigating Gender Bias in English to Romanian Machine Translation

arXiv:2608.08606v1 Announce Type: cross Abstract: Machine translation (MT) systems often fail to correctly translate gender, especially when converting from a gender-neutral language like English to a

MonitorBench: A Comprehensive Benchmark for Chain-of-Thought Monitorability in Large Language Models

arXiv:2603.28590v3 Announce Type: replace Abstract: Large language models (LLMs) can generate chains of thought (CoTs) that are not always causally responsible for their final outputs. When such a mis

Hardware

Nvidia Nemo Switchyard

https://github.com/NVIDIA-NeMo/Switchyard Finally an open source LLM router. An alternative to openrouter fusion and Sakana Fugu. Doesn't look like it does exactly what Sakana Fugu does according to i

hardwarer-localllama

Query-Only Backdoor Attacks on Self-Evolving Skills via Trajectory Poisoning

arXiv:2608.08303v1 Announce Type: new Abstract: Agentic skills improve large language model (LLM) agents by encoding reusable procedures for complex tasks. However, manually authored skills often adap

RAG-Based Auto-Configuration for Industrial Fieldbus Devices

arXiv:2608.08618v1 Announce Type: cross Abstract: Industrial device commissioning requires engineers to manually extract hundreds of protocol-specific parameters from heterogeneous PDF manuals and tra

ReMIND: Orchestrating Modular Large Language Models for Controllable Serendipity A REM-Inspired System Design for Emergent Creative Ideation

arXiv:2601.07121v3 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used not only for problem solving but also for creative ideation; however, generating ideas that

researcharxiv-cs-ai

RippleKV: Cross-Layer KV Cache Allocation via Perturbation Propagation

arXiv:2608.08684v1 Announce Type: cross Abstract: Long-context LLM inference is bottlenecked by KV cache memory, yet distributing a limited cache budget across layers remains challenging. Existing met

Safety Cost of Steering Vectors Is Separable and Reducible

arXiv:2608.08383v1 Announce Type: new Abstract: Steering vectors are a lightweight tool for controlling LLM behavior. However, emerging evidence shows that steering vectors can unintentionally comprom

safetyarxiv-cs-cl

SkillsMetric: Mapping the Detection Boundary of Static Analysis for Malicious Agent Skills

arXiv:2608.08468v1 Announce Type: cross Abstract: Agent Skills---structured packages of instructions and scripts that augment LLM-based agents---are rapidly proliferating, yet their security propertie

Tied Trit-Planes: Constraining PTQTP to a Uniform Nine-Level Quantizer, with a Persistent Folded Format for Disk-Streamed Mixture-of-Experts Serving

arXiv:2608.08910v1 Announce Type: new Abstract: PTQTP decomposes LLM weight matrices into two ternary (trit) planes with two free per-group scales. Tying the scales to a fixed ratio of three collapses

Toward Metacognitive One-Shot Indirect Prompt Injection: Strategy Abstraction Via Outcome-Conditioned Reflection

arXiv:2608.08795v1 Announce Type: cross Abstract: Tool-using large language model (LLM) agents are vulnerable to indirect prompt injection (IPI), in which malicious instructions embedded in external o

What Keeps Agent Skills from Being Reusable? Evidence from 138K SKILL.md Files

arXiv:2608.08453v1 Announce Type: new Abstract: Under the current standard, Agent Skills are SKILL.md files that combine instructions with supporting resources, enabling Large Language Model (LLM) age

Yesterday's Shield, Today's Spear: A Self-Evolving Safety Guardrail in Production

arXiv:2608.08471v1 Announce Type: new Abstract: Deployed LLM safety guardrails are predominantly static: trained once and frozen at release, while new jailbreak techniques and previously un-addressed

A Multi-Agent Framework for Automated Coarse-Grained Molecular Dynamics of Polymers

arXiv:2608.06694v1 Announce Type: new Abstract: Coarse-grained (CG) molecular dynamics extends polymer simulation beyond the scales accessible to all-atom (AA) methods, but bottom-up CG modeling is la

10 Aug 2026

A Picture is Worth a Thousand Tokens: How Vision Language Models Cut AI Energy Costs While Improving Accuracy

arXiv:2608.07427v1 Announce Type: new Abstract: LLM inference accounts for over 90% of AI operational energy, scaling directly with input token count---a critical inefficiency for telecom network anal