Search: “llm”

RouteNLP: Closed-Loop LLM Routing with Conformal Cascading and Distillation Co-Optimization

arXiv:2604.23577v1 Announce Type: new Abstract: Serving diverse NLP workloads with large language models is costly: at one enterprise partner, inference costs exceeded $200K/month despite over 70% of

28 Apr 2026

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Synthetic Eggs in Many Baskets: The Impact of Synthetic Data Diversity on LLM Fine-Tuning

DGX agent

arXiv:2511.01490v2 Announce Type: replace Abstract: As synthetic data becomes widely used in language model development, understanding its impact on model behavior is crucial. This paper investigates

safetyarxiv-cs-cl

28 Apr 2026

TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Training

arXiv:2604.24088v1 Announce Type: cross Abstract: Handling communication overhead in large-scale tensor-parallel training remains a critical challenge due to the dense, near-zero distributions of inte

28 Apr 2026

When Corrective Hints Hurt: Prompt Design in Reasoner-Guided Repair of LLM Overcaution on Entailed Negations under OWL~2~DL

arXiv:2604.23398v1 Announce Type: new Abstract: We report a reproducible error pattern in GPT-5.4 on OWL~2~DL compliance queries: the model frequently answers ``unknown'' when the reasoner-entailed an

28 Apr 2026

Dharma, Data and Deception: An LLM-Powered Rhetorical Analysis of Cow-Urine Health Claims on YouTube

arXiv:2604.22606v1 Announce Type: new Abstract: Health misinformation remains one of the most pressing challenges on social media, particularly when cultural traditions intersect with scientific-sound

PermaFrost-Attack: Stealth Pretraining Seeding(SPS) for planting Logic Landmines During LLM Training

arXiv:2604.22117v1 Announce Type: cross Abstract: Aligned large language models(LLMs) remain vulnerable to adversarial manipulation, and their dependence on web-scale pretraining creates a subtle but

safetyarxiv-cs-ai

SSG: Logit-Balanced Vocabulary Partitioning for LLM Watermarking

arXiv:2604.22438v1 Announce Type: cross Abstract: Watermarking has emerged as a promising technique for tracing the authorship of content generated by large language models (LLMs). Among existing appr

The Shape of Adversarial Influence: Characterizing LLM Latent Spaces with Persistent Homology

arXiv:2505.20435v3 Announce Type: replace-cross Abstract: Existing interpretability methods for Large Language Models (LLMs) predominantly capture linear directions or isolated features. This overlook

FairyFuse: Multiplication-Free LLM Inference on CPUs via Fused Ternary Kernels

arXiv:2604.20913v1 Announce Type: new Abstract: Large language models are increasingly deployed on CPU-only platforms where memory bandwidth is the primary bottleneck for autoregressive generation. We

Grounding Machine Creativity in Game Design Knowledge Representations: Empirical Probing of LLM-Based Executable Synthesis of Goal Playable Patterns under Structural Constraints

arXiv:2603.07101v3 Announce Type: replace Abstract: Creatively translating complex gameplay ideas into executable artifacts (e.g., games as Unity projects and code) remains a central challenge in comp

MCAP: Deployment-Time Layer Profiling for Memory-Constrained LLM Inference

arXiv:2604.21026v1 Announce Type: new Abstract: Deploying large language models to heterogeneous hardware is often constrained by memory, not compute. We introduce MCAP (Monte Carlo Activation Profili

Propensity Inference: Environmental Contributors to LLM Behaviour

arXiv:2604.21098v1 Announce Type: new Abstract: Motivated by loss of control risks from misaligned AI systems, we develop and apply methods for measuring language models' propensity for unsanctioned b

Separable Expert Architecture: Toward Privacy-Preserving LLM Personalization via Composable Adapters and Deletable User Proxies

arXiv:2604.21571v1 Announce Type: new Abstract: Current model training approaches incorporate user information directly into shared weights, making individual data removal computationally infeasible w

Are LLM Uncertainty and Correctness Encoded by the Same Features? A Functional Dissociation via Sparse Autoencoders

arXiv:2604.19974v1 Announce Type: cross Abstract: Large language models can be uncertain yet correct, or confident yet wrong, raising the question of whether their output-level uncertainty and their a

Breaking the Assistant Mold: Modeling Behavioral Variation in LLM Based Procedural Character Generation

arXiv:2601.03396v2 Announce Type: replace Abstract: Procedural content generation has enabled vast virtual worlds through levels, maps, and quests, but large-scale character generation remains underex

safetyarxiv-cs-cl

From Signal Degradation to Computation Collapse: Uncovering the Two Failure Modes of LLM Quantization

arXiv:2604.19884v1 Announce Type: cross Abstract: Post-Training Quantization (PTQ) is critical for the efficient deployment of Large Language Models (LLMs). While 4-bit quantization is widely regarded

How Much Does Persuasion Strategy Matter? LLM-Annotated Evidence from Charitable Donation Dialogues

arXiv:2604.19783v1 Announce Type: new Abstract: Which persuasion strategies, if any, are associated with donation compliance? Answering this requires fine-grained strategy labels across a full corpus

Knapsack Optimization-based Schema Linking for LLM-based Text-to-SQL Generation

arXiv:2502.12911v3 Announce Type: replace Abstract: Generating SQLs from user queries is a long-standing challenge, where the accuracy of initial schema linking significantly impacts subsequent SQL ge

MGDA-Decoupled: Geometry-Aware Multi-Objective Optimisation for DPO-based LLM Alignment

arXiv:2604.20685v1 Announce Type: new Abstract: Aligning large language models (LLMs) to desirable human values requires balancing multiple, potentially conflicting objectives such as helpfulness, tru

safetyarxiv-cs-lg

MixLLM: LLM Quantization with Global Mixed-precision between Output-features and Highly-efficient System Design

arXiv:2412.14590v2 Announce Type: replace Abstract: Quantization has become one of the most effective methodologies to compress LLMs into smaller size. However, the existing quantization solutions sti

QuantaAlpha: An Evolutionary Framework for LLM-Driven Alpha Mining

arXiv:2602.07085v2 Announce Type: replace-cross Abstract: Financial markets are noisy and non-stationary, making alpha mining highly sensitive to noise in backtesting results and sudden market regime

The Model Says Walk: How Surface Heuristics Override Implicit Constraints in LLM Reasoning

arXiv:2603.29025v2 Announce Type: replace-cross Abstract: Large language models systematically fail when a salient surface cue conflicts with an unstated feasibility constraint. We study this through

The Tool-Overuse Illusion: Why Does LLM Prefer External Tools over Internal Knowledge?

arXiv:2604.19749v1 Announce Type: new Abstract: Equipping LLMs with external tools effectively addresses internal reasoning limitations. However, it introduces a critical yet under-explored phenomenon

safetyarxiv-cs-ai

Discovering a Shared Logical Subspace: Steering LLM Logical Reasoning via Alignment of Natural-Language and Symbolic Views

arXiv:2604.19716v1 Announce Type: new Abstract: Large Language Models (LLMs) still struggle with multi-step logical reasoning. Existing approaches either purely refine the reasoning chain in natural l

safetyarxiv-cs-cl

Harmful Intent as a Geometrically Recoverable Feature of LLM Residual Streams

arXiv:2604.18901v1 Announce Type: cross Abstract: Harmful intent is geometrically recoverable from large language model residual streams: as a linear direction in most layers, and as angular deviation

Less Is More: Cognitive Load and the Single-Prompt Ceiling in LLM Mathematical Reasoning

arXiv:2604.18897v1 Announce Type: new Abstract: We present a systematic empirical study of prompt engineering for formal mathematical reasoning in the context of the SAIR Equational Theories Stage 1 c

LLM-as-Judge Framework for Evaluating Tone-Induced Hallucination in Vision-Language Models

arXiv:2604.18803v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) are increasingly deployed in settings where reliable visual grounding carries operational consequences, yet their behavi

Mind the Unseen Mass: Unmasking LLM Hallucinations via Soft-Hybrid Alphabet Estimation

arXiv:2604.19162v1 Announce Type: new Abstract: This paper studies uncertainty quantification for large language models (LLMs) under black-box access, where only a small number of responses can be sam

ODMA: On-Demand Memory Allocation Strategy for LLM Serving on LPDDR-Class Accelerators

arXiv:2512.09427v5 Announce Type: replace-cross Abstract: Existing memory management techniques severely hinder efficient Large Language Model serving on accelerators constrained by poor random-access

Reinforcement Learning Improves LLM Accuracy and Reasoning in Disease Classification from Radiology Reports

arXiv:2604.19060v1 Announce Type: new Abstract: Accurate disease classification from radiology reports is essential for many applications. While supervised fine-tuning (SFT) of lightweight LLMs improv

safetyarxiv-cs-ai

Right for the Wrong Reasons: Epistemic Regret Minimization for LLM Causal Reasoning

arXiv:2602.11675v3 Announce Type: replace Abstract: Large language models may answer causal questions correctly for the wrong reasons, substituting associational shortcuts P(Y|X) for the interventiona

Two-dimensional early exit optimisation of LLM inference

arXiv:2604.18592v1 Announce Type: cross Abstract: We introduce a two-dimensional (2D) early exit strategy that coordinates layer-wise and sentence-wise exiting for classification tasks in large langua

Unlocking the Edge deployment and ondevice acceleration of multi-LoRA enabled one-for-all foundational LLM

arXiv:2604.18655v1 Announce Type: cross Abstract: Deploying large language models (LLMs) on smartphones poses significant engineering challenges due to stringent constraints on memory, latency, and ru

Tutorials

A Practical Guide to LLM Fine Tuning

This guide from Databricks covers practical techniques and best practices for fine-tuning large language models, including methodologies for adapting pre-trained LLMs to specific tasks and domains. It

tutorialsdatabricks

AeroRAG: Structured Multimodal Retrieval-Augmented LLM for Fine-Grained Aerial Visual Reasoning

model-releasesarxiv-cs-cv

arXiv:2604.17889v1 Announce Type: new Abstract: Despite recent progress in multimodal large language models (MLLMs), reliable visual question answering in aerial scenes remains challenging. In such sc

Are Emotion and Rhetoric Neurons in LLM? Neuron Recognition and Adaptive Masking for Emotion-Rhetoric Prediction Steering

arXiv:2604.17255v1 Announce Type: new Abstract: Accurate comprehension and controllable generation of emotion and rhetoric are pivotal for enhancing the reasoning capabilities of large language models

Beyond 'I Don't Know': Evaluating LLM Self-Awareness in Discriminating Data and Model Uncertainty

arXiv:2604.17293v1 Announce Type: new Abstract: Reliable Large Language Models (LLMs) should abstain when confidence is insufficient. However, prior studies often treat refusal as a generic 'I don't k

Beyond Reproduction: A Paired-Task Framework for Assessing LLM Comprehension and Creativity in Literary Translation

arXiv:2604.18169v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used for creative tasks such as literary translation. Yet translational creativity remains underexplored a

Bielik Guard: Efficient Polish Language Safety Classifiers for LLM Content Moderation

arXiv:2602.07954v4 Announce Type: replace Abstract: As Large Language Models (LLMs) become increasingly deployed in Polish language applications, the need for efficient and accurate content safety cla

CAPC-CG: A Large-Scale, Expert-Directed LLM-Annotated Corpus of Adaptive Policy Communication in China

arXiv:2510.08986v2 Announce Type: replace Abstract: We introduce CAPC-CG, the Chinese Adaptive Policy Communication (Central Government) Corpus, the first open dataset of Chinese policy directives ann

safetyarxiv-cs-cl

CoAct: Co-Active LLM Preference Learning with Human-AI Synergy

arXiv:2604.17501v1 Announce Type: new Abstract: Learning from preference-based feedback has become an effective approach for aligning LLMs across diverse tasks. However, high-quality human-annotated p

Context Matters: Peer-Aware Student Behavioral Engagement Measurement via VLM Action Parsing and LLM Sequence Classification

arXiv:2601.06394v3 Announce Type: replace Abstract: Understanding student behavior in the classroom is essential to improve both pedagogical quality and student engagement. Existing methods for predic

researcharxiv-cs-cv

Agents

Diversity Collapse in Multi-Agent LLM Systems: Structural Coupling and Collective Failure in Open-Ended Idea Generation

arXiv:2604.18005v1 Announce Type: cross Abstract: Multi-agent systems (MAS) are increasingly used for open-ended idea generation, driven by the expectation that collective interaction will broaden the

agentsarxiv-cs-cl

Guardrails in Logit Space: Safety Token Regularization for LLM Alignment

arXiv:2604.17210v1 Announce Type: new Abstract: Fine-tuning well-aligned large language models (LLMs) on new domains often degrades their safety alignment, even when using benign datasets. Existing sa

Improving LLM Code Reasoning via Semantic Equivalence Self-Play with Formal Verification

arXiv:2604.17010v1 Announce Type: new Abstract: We introduce a self-play framework for semantic equivalence in Haskell, utilizing formal verification to guide adversarial training between a generator

Knowing When to Quit: A Principled Framework for Dynamic Abstention in LLM Reasoning

arXiv:2604.18419v1 Announce Type: cross Abstract: Large language models (LLMs) using chain-of-thought reasoning often waste substantial compute by producing long, incorrect responses. Abstention can m

LiveFact: A Dynamic, Time-Aware Benchmark for LLM-Driven Fake News Detection

arXiv:2604.04815v2 Announce Type: replace Abstract: The rapid development of Large Language Models (LLMs) has transformed fake news detection and fact-checking tasks from simple classification to comp

LLM Hypnosis: Exploiting User Feedback for Unauthorized Knowledge Injection to All Users

arXiv:2507.02850v3 Announce Type: replace Abstract: We describe a vulnerability in language models (LMs) trained with user feedback, whereby a single user can persistently alter LM knowledge and behav