Search: “llm”

Probability Contracts: Accuracy, Coherence, and Decisions Across LLM Interfaces

arXiv:2609.37470v1 Announce Type: new Abstract: A probability used for a decision should refer to the same event across equivalent requests. We introduce probability contracts, a benchmark connecting

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Risk-Controlled Selective LLM Answering by Pricing Label-Free Checks

DGX agent

arXiv:2609.37493v1 Announce Type: cross Abstract: Serving an answer from a large language model requires deciding when to abstain, yet a verifier's ranking accuracy alone does not determine the error

researcharxiv-cs-ai

30 Sept 2026

SPLASH: Switching Parallel Layouts of Attention with Seamless Handoff for LLM Serving

arXiv:2609.37626v1 Announce Type: cross Abstract: No single way of parallelizing attention serves large language models well under all loads. Low concurrency favors tensor parallelism, many independen

The Default Trap: Rethinking Plan Evaluation in Tool-Using LLM Agents

arXiv:2609.36829v1 Announce Type: new Abstract: An executor can respond strongly to a change in a supplied plan's priority while showing a small change in the same information-selection probability wh

researcharxiv-cs-ai

Trajectory Soup: Pushing the Compute-Scaling Frontier of LLM Mid-training via Diverse Trajectories

arXiv:2609.37169v1 Announce Type: cross Abstract: Mid-training equips pretrained large language models with specialized and reasoning capabilities, but the returns of this stage are bounded since addi

UserProxyBench: Evaluating LLM User Simulators for Agent Benchmarks and Training

arXiv:2609.38043v1 Announce Type: new Abstract: Interactive agent benchmarks and multi-turn reinforcement learning increasingly place a second language model in the role of the user. This simulated us

Using AFM 3 PCC as an alternative to Local LLM

Apple Foundation Model 3 through Private Cloud Compute is directly available in macOS 27.2 through familiar AI clients. For Mac users running local models such as Qwen 3.x or Gemma 4, AFM 3 PCC is a c

model-releasesr-ollama

When Updating Stops Being Learning: Rethinking LLM Self-Evolution via learnable information gain

arXiv:2609.36535v1 Announce Type: new Abstract: Self-evolution lets large language models (LLMs) improve iteratively using their own generated data, but often suffers from self-evolution degeneration:

researcharxiv-cs-cl

Where Does Staleness Accumulate? Pool Aware Effective Staleness Control for Asynchronous RL in LLM Post-Training

arXiv:2609.36830v1 Announce Type: new Abstract: Fully asynchronous reinforcement learning (RL) improves resource utilization in large language model post-training by overlapping rollout generation wit

400+ LLM agents living in a 2004-era MMO server, all local on Qwen3-4B

I wanted a fun way to visualize and stress-test a multi-agent ecosystem, so I pointed it at my favorite MMO (a vanilla-era private server emulator). Live view: https://design.kristiantalley.com/projec

model-releasesr-ollama

AG-CoT: Verified Algorithmic Traces for LLM Program Synthesis on Clifford Circuits

arXiv:2609.33192v1 Announce Type: new Abstract: Scientific code generation can produce executable programs that fail to compute the intended scientific object. We study this problem in language-model

researcharxiv-cs-lg

An RL View of OPD: Least Square Policy Distillation for Sample-Efficient LLM Reasoning

arXiv:2609.35505v1 Announce Type: new Abstract: We study on-policy distillation (OPD) through the lens of reinforcement learning, establishing a connection between the reverse-KL objective in OPD and

Curating Merchant-Matching Training Data with Two Confidence-Gated Local LLM Judges

arXiv:2609.33878v1 Announce Type: cross Abstract: Merchant matching resolves a noisy payment descriptor to a retrieved merchant entity or returns no match. A key challenge in curating training labels

EffiPair: Improving the Efficiency of LLM-generated Code with Differential Execution Feedback

arXiv:2604.05137v2 Announce Type: replace-cross Abstract: Large language models (LLMs) can generate functionally correct programs that differ substantially in execution efficiency. Existing inference-

FORGE: Form-Optimal Routing of Grounded Evidence for Frozen LLM Agents

arXiv:2609.34358v1 Announce Type: new Abstract: In agentic AI systems, frozen foundation models are increasingly deployed as closed-weight API endpoints, making downstream adaptation possible only thr

Frozen Judges, Moving Agents: Version-Dependent LLM-Judge Error and the Limits of Judge-Assisted Agent Evaluation

arXiv:2609.34198v1 Announce Type: new Abstract: Language-model judges compare agent upgrades with their predecessors, but a fixed judge can make version-dependent mistakes. We analyze 35 public coding

agentsarxiv-cs-lg

GroupMask: Layer-Adaptive Group-wise Sparsity for Semi-Structured LLM Pruning

arXiv:2609.33977v1 Announce Type: new Abstract: Semi-structured pruning compresses large language models (LLMs) while keeping a regular sparse structure, but the prevailing N:M pattern fixes the same

Inspector: Conversational and Lightweight Analyzer of Analog Circuit Layouts Using LLM and CNNs

arXiv:2609.34976v1 Announce Type: cross Abstract: The integration of artificial intelligence into computer-aided design frameworks has sparked a shift in the design of analog integrated circuits (ICs)

researcharxiv-cs-cv

LLM Unlearning Evaluation with TRIAGE

arXiv:2609.32103v1 Announce Type: new Abstract: Large language models can memorize private or harmful information, motivating machine unlearning methods that remove targeted knowledge while preserving

Local Ai

OmniTide: Co-Designing Algorithms and Systems for Efficient On-Device Omni-LLM Streaming

arXiv:2609.34653v1 Announce Type: cross Abstract: On-device streaming omni-modal inference safeguards user privacy and eliminates prohibitive per-token API costs, but faces a critical bottleneck: the

local-aiarxiv-cs-lg

Pre-registered tests of solid-state-physics-inspired LLM compression: a cluster-level negative result at small-language-model scale

arXiv:2609.34292v1 Announce Type: cross Abstract: We report a three-month autonomous research-agent program testing five solid-state-physics-inspired compression mappings on pretrained language models

agentsarxiv-cs-lg

Recursive LLM Degradation in Biomedical Question Answering: A Cross-Generation Study

arXiv:2609.34257v1 Announce Type: cross Abstract: Repeatedly training language models on their own generated data may create a synthetic-data feedback loop in which errors and distributional biases ar

researcharxiv-cs-lg

Rethinking Training-Inference Mismatch in LLM Reinforcement Learning: Where It Arises and How to Correct It

arXiv:2609.32444v1 Announce Type: new Abstract: We study training-inference mismatch in reinforcement learning with verifiable rewards (RLVR) for large language models, where rollouts are sampled by a

Share-Borne AI Virus: Memory-Hopping Attacks Across LLM Agents

arXiv:2609.35576v1 Announce Type: cross Abstract: Large language models are increasingly deployed as stateful assistants that retain information across interactions and use tools to read, modify, and

SoFT: Soft Targets for Generalizable LLM Fine-Tuning

arXiv:2609.32493v1 Announce Type: new Abstract: Distillation enables student language models to acquire new capabilities from expert teachers. However, integrating knowledge from multi-teacher, multi-

agentsarxiv-cs-lg

Surprising Success, Repeated Failure: Entropy-Guided Credit Assignment for Exploration in LLM Reasoning

arXiv:2609.33781v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) enhances reasoning in large language models (LLMs) through outcome-level feedback, yet recent appr

Teach to Learn: Hint Annealing for Self-improving LLM Reasoning

arXiv:2609.34975v1 Announce Type: new Abstract: Group Relative Policy Optimization (GRPO) improves language-model reasoning by comparing verified rewards among multiple solution rollouts for each quer

Theory of Scene: Breaking the Symmetry Trap in Multi-Agent LLM Coordination

arXiv:2609.32939v1 Announce Type: new Abstract: Multi-agent systems built on large language models (LLMs) are largely homogeneous, as their agents behave alike even across distinct LLMs. We show that

TRACE: Trajectory-Based Safety Patch Learning for LLM Post-Training Realignment

arXiv:2607.16242v2 Announce Type: replace Abstract: Fine-Tuning-as-a-Service (FTaaS) platforms let users perform supervised fine-tuning (SFT) on customized data, but this pipeline can erode model safe

When Do Model Internals Help? Exploring the Role of Representation Engineering in LLM Safety

arXiv:2609.34771v1 Announce Type: cross Abstract: Reliable AI safeguards require both control mechanisms that reduce unsafe behavior and monitoring mechanisms that detect safety risks during model int

Where Activation Sparsity and KV-Cache Sparsity Cross in LLM Decoding

arXiv:2609.33889v1 Announce Type: new Abstract: At each step, decoding one sequence with a large language model rereads the projection weights, whose traffic is fixed, and the key-value (KV) cache, wh

All In Good Time: Causality-Aware Framework for LLM-Based Simultaneous Speech-to-Speech Translation

arXiv:2609.30416v1 Announce Type: new Abstract: Large Language Models (LLMs) have shown strong performance in low-resource offline translation; however, extending them to simultaneous speech-to-speech

safetyarxiv-cs-cl

AuthGuard-R: Safety-Compliant Mission Hijacking and Dual-Gate Defense for LLM-Controlled Robots

model-releasesarxiv-cs-ro

arXiv:2609.31110v1 Announce Type: new Abstract: Large language models are increasingly used as high-level planners for mobile robots, robot manipulators, and autonomous vehicles. Recent studies show t

CognitiveReality: Robot-Agnostic Semantic Gaussian Mapping with an LLM Agent for Immersive Collaborative VR Teleoperation

arXiv:2609.31418v1 Announce Type: new Abstract: A photorealistic 3D view tells a teleoperator where a robot is, but not what the scene contains, how well each object has been observed, or how to turn

agentsarxiv-cs-ro

Detecting Time Series Anomalies Like an Expert: A Multi-Agent LLM Framework with Specialized Analyzers

arXiv:2605.05725v2 Announce Type: replace Abstract: Time-series anomaly detection often returns scores or intervals, while analysts need to understand the abnormal behavior and the evidence supporting

Local Ai

G^2PTQ: Improving LLM Post-Training Quantization with Generalized Gradient Compensation

arXiv:2609.31009v1 Announce Type: cross Abstract: Post-training quantization (PTQ) is a practical approach to reducing the memory and computational footprint of large language models (LLMs) without re

local-aiarxiv-cs-ai

Game Arena: Strategic LLM Evaluation in Competitive Environments

arXiv:2609.31473v1 Announce Type: new Abstract: We introduce Kaggle Game Arena, an open and ever-expanding platform to evaluate large language models (LLMs) through competitive games. Different from s

researcharxiv-cs-ai

Thinking Less to Simulate Better: Intuitive Prompting Improves LLM Agents Simulating Individual Social Media Reactions, Including Unfamiliar Content

arXiv:2609.30563v1 Announce Type: new Abstract: Platform policies are increasingly tested on artificial users, making agent fidelity important. Yet convincing fake profiles could also manipulate perce

We released VeriLoop E2 (27B, Apache-2.0). The design question behind it: should an LLM be allowed to commit its own state?

model-releasesr-localllama

Disclosure: I’m one of the authors of VeriLoop E2, a 27B model post-trained from Qwen3.8-27B. The weights are Apache-2.0; the harness we evaluate it in is not open source (details at the bottom). This

27 Sept 2026

Local Ai

At what point do LLMs start bootstrapping themselves and generating the next LLM

I suppose technically if that happens it will signal the true start of a singularity because from that point on progress will be exponential and not dependant on humans. right now you still need huge

local-air-localllama

26 Sept 2026

A Wrong Turn Does Not Ruin the Journey: Deviation-Guided Skill Self-Evolution for LLM Agents

arXiv:2609.29154v1 Announce Type: new Abstract: Large language model agents increasingly rely on natural-language skills to solve complex tool-use tasks. However, such tasks often admit multiple valid

Calibrated Decision Models for Autonomous Penetration-Testing Harnesses: JEV and Laya as System One Decision Layers for LLM-Driven Pentest Agents

arXiv:2609.28940v1 Announce Type: cross Abstract: Autonomous penetration-testing harnesses use large language models (LLMs) for reconnaissance, exploitation, and reporting, but often rely on those sam

ChunkRank: Model-Aware Text Chunking and Abstention-Aware Answer Selection for LLM Pipelines

model-releasesarxiv-cs-cl

arXiv:2609.29828v1 Announce Type: new Abstract: We present ChunkRank, an open-source Python library that derives chunk boundaries from a target model's tokenizer and context window, and selects an ans

Decoupling Knowledge and Privacy: Post-Task Self-Distillation Replay for LLM Continual Learning

arXiv:2609.29711v1 Announce Type: cross Abstract: Privacy-preserving continual learning (PPCL) must reduce the reproduction of sensitive content while retaining useful knowledge across sequential task

Encoded but Not Decoded: Layer-Localized Evidence for a Three-Level Gap in LLM Syntax

model-releasesarxiv-cs-cl

arXiv:2609.29848v1 Announce Type: new Abstract: A language model can fail a syntactic test in two distinct ways: by not encoding the relevant structure, or by encoding it but failing to use it at the

Fair Like Us? Auditing LLM Alignment in Resource Allocation

arXiv:2609.29692v1 Announce Type: new Abstract: Fair allocation of scarce, indivisible resources is an important challenge in many societal problems. While there are several formal theories of fairnes

safetyarxiv-cs-ai

LLM Forensics: Where Do Backdoors Hide? Localizing and Controlling Trigger Mechanisms with Sparse Autoencoders

arXiv:2609.07746v2 Announce Type: replace Abstract: Even though backdoors in LLMs have been a growing concern, their inner workings are still under heavy scrutiny. Trigger-based backdoors are easy to

researcharxiv-cs-cl

LLM on a budget part 2, from P102-100 to CMP 50HX.

model-releasesr-localllama

I finally got around to upgrade the GPU's. First a word of warning, when upgrading GPU's on P520 you have to be extra careful not to slot the card on any angle other than straight when installing or w