Search: “open-weights”

From Triage to Impact: A Multi-Hop Reasoning Benchmark for Explainable Persian Hate Speech

arXiv:2610.20370v1 Announce Type: new Abstract: Hate speech has become increasingly prevalent on social media, raising concerns about its detection, interpretation, and moderation. Persian, a relative

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

Geometric Alignment without Functional Equivalence: A Layer-wise Analysis of the Speech-Text Modality Gap

DGX agent

arXiv:2603.01502v2 Announce Type: replace Abstract: End-to-end speech large language models (LLMs) often answer the same question less accurately from speech than from text, even when the spoken conte

safetyarxiv-cs-cl

15 Oct 2026

Glasseek: A Transparent Recipe for Training and Trustworthy Evaluation of DeepSearch Agents

arXiv:2610.16998v1 Announce Type: new Abstract: Deep search enables language agents to solve complex information-seeking tasks through iterative web exploration. However, developing robust open-source

Humans vs. Machines: A Comparison With AI Agents in Software Reverse Engineering

arXiv:2610.14960v1 Announce Type: cross Abstract: Can an autonomous AI agent reverse engineer a binary as well as a skilled human analyst? Do they adopt the same strategy, and can agents scale to real

Instruction-Conditioned Motion Control for Robotic Arms: Language-to-Code Decomposition and Reinforcement Learning

model-releasesarxiv-cs-ro

arXiv:2610.19617v1 Announce Type: new Abstract: Instruction-Conditioned Motion Control (ICMC) separates parameter specification, instruction generation, and code synthesis for language-driven robot co

Lightweight but Linguistically Fair? Quantisation Robustness across Bavarian, German, and English

arXiv:2610.18343v1 Announce Type: cross Abstract: Post-training quantisation reduces the memory cost of large language models, but it remains unclear whether compression preserves language skills of r

LLM Alignment--Utility Asymmetry under Semantic-Preserving Transformations

arXiv:2609.32717v2 Announce Type: replace-cross Abstract: Large Language Model (LLM) alignment is intended to ensure that models remain helpful and safe, but its stability under input distributional s

LLM-jp-4 Technical Report

arXiv:2610.19331v1 Announce Type: new Abstract: This paper presents LLM-jp-4, a family of Japanese sovereign large language models (LLMs) developed with an emphasis on transparency, reproducibility, a

Loose Lips: Understanding Privacy Leakage in Agentic Dialogue

arXiv:2610.14071v1 Announce Type: cross Abstract: As LLM agents are increasingly deployed in high-stakes enterprise settings, understanding how they handle sensitive information in agentic dialogue ha

MA-Bench: Measuring Operational Planning in Language Models

arXiv:2610.19138v1 Announce Type: cross Abstract: How well can a language model turn a large, incomplete and partly contradictory body of documents into a plan when no single correct answer exists? We

Agents

Misinformation Propagation in Benign Multi-Agent Systems

arXiv:2606.16710v2 Announce Type: replace-cross Abstract: Multi-agent systems (MAS) are credited with solving complex problems, yet their agents routinely act on context that may be false. Prior work

agentsarxiv-cs-cl

Pre-Trained Language Models as Zero-Shot Tools for Social Science Research

arXiv:2610.19048v1 Announce Type: new Abstract: Record linkage and demographic inference from names are routine demands in social science. Most approaches to these problems require labeled training da

safetyarxiv-cs-cl

RCCM: Diagnosing Gains and Harms in Reasoning-Prompt Transfer

arXiv:2610.19872v1 Announce Type: new Abstract: A reasoning prompt can improve one deployment and degrade another. RCCM (Reasoning-Control Capability Mismatch) examines these reversals by pairing answ

REPRO: Reflective Evolutionary Prompt Optimizer for Automated LLM Red-Teaming

arXiv:2610.15513v1 Announce Type: cross Abstract: Safety-aligned large language models (LLMs) remain vulnerable to adversarial prompts that bypass their safety training. Existing automated red-teaming

Review Rarely Beats the Stronger Model: Separating Verification from Capability Transfer in LLM Pipelines

arXiv:2610.15897v2 Announce Type: cross Abstract: Qwen3-8B reviewing Llama-3.1-8B in math adds 33.2 points, yet Qwen3-8B alone is 6.4 points more accurate than the result. Reviewing another model's an

Running Fewer Loops Removes Refusal Behaviour in Recurrent-Depth Language Models

arXiv:2610.15218v1 Announce Type: new Abstract: A recurrent-depth language model reuses one block of weights several times in a row. The number of repetitions, which we call loops, is chosen when the

SEA-HELM Audio: Evaluating Multilingual ASR-LLM Pipeline Performance Across SEA-HELM Subsets

arXiv:2610.19869v1 Announce Type: cross Abstract: We introduce SEA-HELM Audio, a Southeast Asian (SEA) multilingual evaluation framework that extends the text-based SEA-HELM benchmark to assess perfor

SLS-Bench: A Benchmark for Incident Log Summarization with Synthetic Observability Data

arXiv:2610.16412v1 Announce Type: new Abstract: As IT systems grow in complexity, automated agent-driven incident response becomes increasingly critical. A key capability for these agents is summarizi

SolBehave: Evaluating the Behavioural Correctness of LLM-Generated Smart Contracts

arXiv:2610.18745v1 Announce Type: cross Abstract: Large language models increasingly generate Solidity smart contracts that, once deployed, can never be patched, making a single latent fault permanent

SpecCloak: Probing Harmful Code Generation through Engineering Specifications

arXiv:2610.17907v1 Announce Type: cross Abstract: The code generation capabilities of large language models (LLMs) can equally be misused to implement harmful functionality, underscoring the need to u

safetyarxiv-cs-ai

StartLux-Decision: AutoThink Decision Models with Reinforcement Learning for Calibrated Deliberative Decisions

arXiv:2610.17309v1 Announce Type: cross Abstract: We introduce StartLux-Decision, decision models from 0.8B to 27B parameters that combine efficient single-pass prediction with adaptive, confidence-gu

Steering Vectors for Competing Ethical Frameworks Are Interchangeable: A Competing-Target Test for Activation Steering

arXiv:2610.19359v1 Announce Type: cross Abstract: Activation steering adds a vector to a model's hidden states to control a named behaviour. A gain on that behaviour, however, does not show that the v

TEX-MEX: Similarity-Guided Temporal Contrastive Learning for Time Series

model-releasesarxiv-cs-lg

arXiv:2610.19557v1 Announce Type: new Abstract: Contrastive learning for time series depends on deciding which pairs of inputs should be treated as similar and dissimilar. Because few augmentations re

The Agent Knew: Internal States Forecast Unauthorized Actions and Stop Them, and the Lies That Follow

arXiv:2610.17301v1 Announce Type: new Abstract: Language-model agents sometimes act beyond their authorization and then misreport what they did. On CAPE-Bench, matched agent tasks in which one fact de

When Pleasing the User Compromises Care: AI Sycophancy in Caregiver-Mediated Healthcare

arXiv:2610.14360v1 Announce Type: new Abstract: Large language model (LLM) alignment often assumes that the user interacting with a model is also the primary beneficiary of its behavior. Yet in many c

Agaram 2.5B: The First Tamil-First Language Model Trained from Scratch with Tool Calling. A Playbook for Under-Represented Languages

model-releasesarxiv-cs-lg

arXiv:2610.18420v1 Announce Type: cross Abstract: We trained Agaram 2.5B, a 2.5-billion-parameter Tamil-English language model, from scratch on about 50 billion tokens (58% Tamil, the rest English, ma

Tutorials

CartoVec: General-Purpose Location Embeddings from Open Map Data

arXiv:2610.17190v1 Announce Type: new Abstract: Geographic coordinates are common inputs in applied machine learning, but latitude and longitude describe position rather than the surrounding environme

tutorialsarxiv-cs-lg

Applications

Embedding Graph Kernel Entropy (EGKE): Fine-grained Hallucination Detection for LLMs from Embeddings

arXiv:2610.14214v1 Announce Type: new Abstract: Hallucination detection is critical for ensuring the reliability of Large Language Models (LLMs) in real-world deployments. While recent works, such as

applicationsarxiv-cs-lg

HarnessTTT: Learn to Optimize Harness at Test-Time for Scientific Discovery

model-releasesarxiv-cs-lg

arXiv:2610.17370v1 Announce Type: new Abstract: Scientific discovery with language models requires finding the best valid solution within a finite computational budget. We study how feedback from this

The Gradient That Top-k Distillation Never Computes

arXiv:2610.15589v1 Announce Type: new Abstract: Many deployed distillation pipelines show the student a truncated view of the teacher: keep the top-k probabilities, renormalize, and minimize a diverge

safetyarxiv-cs-lg

Truncation Samplers Are Nested: Comparing Decoding Rules Is a Measurement Problem

model-releasesarxiv-cs-cv

arXiv:2610.15570v1 Announce Type: cross Abstract: Settling one comparison between two truncation samplers has cost six thousand A100-hours, and nothing says in advance how far apart two samplers can b

Unfair and Unsafe: Everyday Bias Triggers Emergent Misalignment

arXiv:2610.15961v1 Announce Type: cross Abstract: Emergent misalignment (EM) occurs when fine-tuning a language model to a narrow task makes it misaligned on unrelated ones. Because organizations now

safetyarxiv-cs-lg

AMLworld-Compact: Importance-Weighted Downsampling for Cost-Effective LLM Evaluation and Error Diagnosis

arXiv:2610.13627v1 Announce Type: new Abstract: Evaluating large language models (LLMs) on AMLworld's smallest test splits is costly because each of over a million targets requires a separate graph pr

Evaluating LLM Creativity through Long-Tail Performance

arXiv:2610.12998v1 Announce Type: new Abstract: Creativity benchmarks for large language models (LLMs) vary widely in task and metric design. This raises the question of whether they capture a common

GeopoliticalQA: Benchmarking Retrieval and Generation over Intelligence Reports

arXiv:2610.12617v1 Announce Type: new Abstract: Geopolitical and strategic intelligence reports constitute a rich source of expert knowledge and have increasing value in the current political climate,

MAGMA-Bench: Evaluating Long-Horizon Embodied Agents under Dynamic Tasks and Execution Failures

arXiv:2610.13534v1 Announce Type: new Abstract: Long-horizon embodied agents must maintain task-relevant information, coordinate heterogeneous robots, and adapt their decisions as execution unfolds. E

MoPrO-SD: Modular Prompt Optimization for Stance Detection

arXiv:2610.13386v1 Announce Type: new Abstract: Stance detection remains challenging for large language models (LLMs) due to implicit opinions, sarcasm, and context-dependent expressions of stance. Al

Task Context Shapes Component Roles in Language Models

arXiv:2610.13895v1 Announce Type: cross Abstract: Large language models reuse the same attention heads across tasks, but identifying these shared heads does not explain what they do differently in eac

Unreliable Narrator Framing as a Cost-effective Mechanism for Satire Explanation of Australian and American Satirical News

arXiv:2610.12863v1 Announce Type: new Abstract: Satirical news is a unique form of misinformation: it is intended to be humorous by the original publication source but can be misinterpreted. Explanati

AgentHorizon: Evaluating Agentic Judges for Long-Horizon Computer-Use Tasks

arXiv:2610.11050v1 Announce Type: new Abstract: Computer-use agents are capable of completing complex tasks, increasing the use of automatic judges to determine success, either for training or for eva

AI4Fire: Evaluating Large Language Models on Wildfire Tasks

arXiv:2610.10946v1 Announce Type: new Abstract: Large language models (LLMs) are entering wildfire management, where overstated evaluations can cost property and lives. How do they perform on wildfire

Caught in the Act: Probes Effectively Detect Sabotage and Catch Unverbalized Deception

arXiv:2610.12445v1 Announce Type: cross Abstract: Recent incidents have highlighted the challenge of monitoring LLM agents and the danger of models deceiving people. We show that white-box deception d

researcharxiv-cs-ai

Applications

Cited but Not Consulted: A Counterfactual Audit of Legal Chain-of-Thought Faithfulness

arXiv:2610.12361v1 Announce Type: new Abstract: Large language models increasingly justify legal decisions by naming the statute or precedent behind a verdict, treated as evidence that the decision fo

applicationsarxiv-cs-ai

Is Memorization Context-Sensitive? Prefix-Based Extraction Beyond Isolated Prefixes

arXiv:2610.12085v1 Announce Type: new Abstract: Large language models (LLMs) can expose memorized training sequences under prefix-based extraction: given a prefix from a training example, the model ma

researcharxiv-cs-ai

Mine Odyssey: Benchmarking Spatial Agentic Intelligence in the Wild

arXiv:2610.11328v1 Announce Type: new Abstract: Advances in foundation models are driving efforts to introduce agents to assist people in the physical world. Such agents require agentic spatial intell

MiniVer-V: Identifying Minimal Sufficient Evidence for Short Video Verification

arXiv:2610.11233v1 Announce Type: cross Abstract: A core challenge in short-video fact-checking is identifying which evidence is sufficient to support a verification conclusion. Existing approaches ei

NanoProof: Open and Efficient Automated Theorem Proving in Lean 4

arXiv:2610.11605v1 Announce Type: cross Abstract: We introduce NanoProof, to our knowledge the first factorized execution-guided theorem prover in Lean 4 whose training data, extraction tooling, train

Narrow and Deep: An Ontology Tower as the Knowledge of an LLM Agent for an Industrial Equipment System

arXiv:2610.11768v1 Announce Type: cross Abstract: Large language model (LLM) agents are beginning to operate industrial energy equipment, and what they get right depends on what they are told about th