safety

GRIP: Grounded Reasoning via Information-Restricted Premises

arXiv:2608.16776v1 Announce Type: new Abstract: High-capacity encoders in retrieval-augmented generation (RAG) can let the query dominate the latent state, leaving retrieved evidence functionally irre

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Grounding LTL Tasks in Sub-Symbolic RL Environments for Zero-Shot Generalization

DGX agent

arXiv:2602.09761v2 Announce Type: replace-cross Abstract: In this work we address the problem of training a Reinforcement Learning agent to follow multiple temporally-extended instructions expressed i

safetyarxiv-cs-ai

18 Aug 2026

Guaranteed Adaptive Modality Acquisition: When the Policy Chooses Its Own Calibration Group

arXiv:2608.15520v1 Announce Type: new Abstract: A multimodal system may begin inference holding only some of its inputs and may acquire the rest at a cost. With adaptive acquisition, the policy determ

HAF: Adapting Generalist VLAs to Humanoid Whole-Body Loco-manipulation via Hierarchical Action Flow and Spectral Latent RL

arXiv:2608.16837v1 Announce Type: cross Abstract: Humanoid robots hold great promise as general-purpose agents in human-centered environments, yet generalist vision-language-action (VLA) foundation mo

Hallucination Span Detection with Input-Side Evidence Alignment

arXiv:2608.15804v1 Announce Type: new Abstract: Hallucinations remain a major obstacle to the reliable use of large language models (LLMs) in conditional text generation. Existing methods primarily as

safetyarxiv-cs-cl

HarmTrace: Anchor-Calibrated Decoupled Optimization for Fine-Grained Target Identification in Harmful Memes

arXiv:2608.16622v1 Announce Type: cross Abstract: Multimodal harmful meme detection is typically formulated as image--text harmfulness classification. A model may correctly predict harmfulness while m

I had a great time discussing with @SciencePetr about the growing risks of AI and how these often stem from current AI training techniques. …

I had a great time discussing with @SciencePetr about the growing risks of AI and how these often stem from current AI training techniques. Thanks for coming to visit @LawZero_ and for covering these

safetyyoshua-bengio--x

I-Perceive: A Foundation Model for Active Perception with Language Instructions

arXiv:2603.00600v2 Announce Type: replace Abstract: Active perception - the ability of a robot to proactively select viewpoints to acquire task-relevant information - is essential for robust operation

I would like to publicly commend @sama and @OpenAI for this taking it at face value. It is very clear that strange and perhaps dangerous thi…

I would like to publicly commend @sama and @OpenAI for this taking it at face value. It is very clear that strange and perhaps dangerous things are happening and our systems are not ready for this. Mo

safetyemad-mostaque--x

Identifying Confusion Trends in Concept-based XAI for Multi-Label Classification

arXiv:2608.15731v1 Announce Type: cross Abstract: Deep Neural Networks (DNNs) deployed in high-risk domains, such as healthcare and autonomous driving, must be not only accurate but also understandabl

If alignment issues are becoming big enough that OpenAI is willing to commit 20% of research inference compute to chain-of-thought monitorin…

If alignment issues are becoming big enough that OpenAI is willing to commit 20% of research inference compute to chain-of-thought monitoring, that suggests that alignment issues are becoming a pretty

safetyethan-mollick--x

Imagining Recovery: Inference-Time Counterfactual Realignment for Vision-Language-Action Models

arXiv:2608.14822v1 Announce Type: cross Abstract: Vision-language-action (VLA) models have improved the flexibility and generality of robotic manipulation, yet they remain fragile to online disruption

Individual Disempowerment through an Advice Channel: Control Loss when Influence is Endogenous

arXiv:2608.14795v1 Announce Type: new Abstract: An AI that can only give advice seems safe: the human is always free to ignore it. That is the premise of the boxing tradition in AI safety, and its lon

Inference-Time Mitigation of Adversarial Political Bias in Large Language Models

arXiv:2608.14629v1 Announce Type: cross Abstract: As Large Language Models (LLMs) become the mainstay for information retrieval and summarization tasks, ensuring that they are always non-partisan and

Inferential Evaluation of Surrogate-Derived Models under Covariate Shift

arXiv:2608.15783v1 Announce Type: cross Abstract: In transfer-learning settings, a model derived from abundant surrogate labels may be deployed in a target population where gold-standard outcomes are

LaGSplat: Inferring Physics-Governed Interactive Simulation from Monocular Video Using Latent Lagrangian Gaussian Splatting

arXiv:2608.16324v1 Announce Type: new Abstract: We present LaGSplat (Latent Lagrangian Gaussian Splatting), a framework that infers interactive, physics-governed dynamics from one or a few monocular v

Le Critique: Privileged Value Functions for LLM Reinforcement Learning

arXiv:2608.16739v1 Announce Type: new Abstract: Reinforcement learning algorithms for Large Language Models (LLMs) are largely distinguished by their variance reduction strategy. Group-relative method

Learning Stock Trading Policies via Barycenter-Based Adversarial Inverse Reinforcement Learning

arXiv:2608.15770v1 Announce Type: new Abstract: Designing effective trading strategies using reinforcement learning remains challenging due to delayed and noisy rewards, poor exploration, and the diff

Learning to Unlearn: Machine Unlearning via Learning the Unlearning Behaviors

arXiv:2608.16700v1 Announce Type: cross Abstract: Various machine unlearning techniques have been developed in response to privacy legislation requirements, enabling individuals to exercise their lega

Left-Branching Transformers Excel at Right-Branching Languages: Data Shapes Word Order Preferences in Language Models

arXiv:2608.15129v1 Announce Type: cross Abstract: We systematically compare word order preferences in decoder-only language models across 192 artificial languages and typologically diverse natural lan

Leveraging Machine Unlearning for Cost-Efficient Preference Alignment

arXiv:2504.06659v2 Announce Type: replace-cross Abstract: Despite advances in Preference Alignment (PA) for Large Language Models (LLMs), mainstream methods like reinforcement learning with human feed

LLM-Based Hierarchical Coordinated Control with Continuation-Aware Policy Learning

arXiv:2608.15041v1 Announce Type: new Abstract: Coordinating multiple interacting units in complex engineering systems is challenging when system interactions are difficult to model, operational infor

LLM Safety Alignment in Low-Resource Languages: A Systematic Literature Review

arXiv:2608.14626v1 Announce Type: cross Abstract: Large Language Models (LLMs) have achieved substantial progress in safety alignment, yet their safety guarantees remain significantly weaker in low-re

Making two action heads agree: coordination mechanisms and a runtime collapse certificate for flow-matching policies

arXiv:2608.15748v1 Announce Type: new Abstract: A dual-representation flow-matching policy decodes each predicted motion into joint and end-effector spaces, and the residual between the two kinematica

Marker-Constrained Pose-Graph Correction for Cross-Platform Georeferencing in GNSS-Denied Environments

arXiv:2608.16281v1 Announce Type: new Abstract: Autonomous operation in GNSS-denied environments requires heterogeneous mapping pipelines to maintain a consistent spatial reference. This paper present

Matched Outcomes, Divergent Gaze: How Foveated MLLMs Search Compared to Humans

arXiv:2608.16514v1 Announce Type: cross Abstract: Human visual search is serial: the fovea must land on a candidate to confirm it, and those landings form a scanpath. Whether multimodal large language

MatchingPolicy: Correspondence-Aware Policy Enables Cross-Object In-Context Learning

arXiv:2608.16715v1 Announce Type: new Abstract: In-context imitation learning enables few-shot policy generalization but struggles to maintain performance on unseen objects and novel scenarios. To add

Max-Q Selective Imitation for Human-in-the-Loop Online Robot Learning

arXiv:2608.15088v1 Announce Type: cross Abstract: Human-in-the-loop (HIL) online reinforcement learning for real robots must absorb human interventions quickly while continuing to improve beyond the h

Measuring Obedience to Authority Across Large Language Models with the Milgram Paradigm

arXiv:2608.16177v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed as agents that operate equipment, execute instructions, and act inside institutional hierarchie

Measuring the Prevalence of Policy Violating Content with ML Assisted Sampling and LLM Labeling

arXiv:2602.18518v2 Announce Type: replace Abstract: Content safety teams need metrics that reflect what users actually experience, not only what is reported. We study prevalence: the fraction of user

MicroVerse: An Instrument for Measuring Self-Authored Identity Drift in Long-Horizon Multi-Agent Language-Model Simulations

arXiv:2608.15844v1 Announce Type: new Abstract: Long-horizon, multi-agent language model (LM) simulations are widely proposed for studying social behavior, yet instruments to measure whether persona-c

safetyarxiv-cs-cl

MINT: Min-Selection Preference Distillation for Balanced Multi-Objective Alignment

arXiv:2608.14828v1 Announce Type: new Abstract: Aligning a language agent to several objectives at once is a persistent failure mode of preference-based training: when objectives are combined additive

MLLM-Guided Semantic Correction for Text-to-Video Generation

arXiv:2608.16513v1 Announce Type: cross Abstract: Recent advances in diffusion models and Transformer architectures have led to significant progress in text-to-video generation. However, these models

MM-BEV: Enhancing Timeliness by Computing Where and When it Matters

arXiv:2608.15437v1 Announce Type: cross Abstract: Multimodal bird's-eye-view (BEV) perception combines LiDAR depth accuracy with dense camera semantics, but its high computational cost and imperfect s

Modality-Invariant Coarse-to-Fine Retinal Image Registration

arXiv:2608.14829v1 Announce Type: cross Abstract: Retinal image registration is essential for ophthalmic diagnosis, longitudinal disease monitoring, and multimodal retinal image analysis. Existing ret

Model Hypnosis: Strong control of AI via additive subliminal effects

arXiv:2608.16834v1 Announce Type: cross Abstract: We demonstrate that AI models are broadly susceptible to a phenomenon we call model hypnosis, in which individually weak and seemingly irrelevant cues

Most parsers treat tracked changes as noise. Result: a deleted clause comes back as live text, and your pipeline reads a document that says …

Most parsers treat tracked changes as noise. Result: a deleted clause comes back as live text, and your pipeline reads a document that says the opposite of what it means. LlamaParse now handles revisi

safetyjerry-liu--x

Multi-Bin Batching for Increasing LLM Inference Throughput

arXiv:2412.04504v2 Announce Type: replace Abstract: As large language models (LLMs) grow in popularity for their diverse capabilities, improving the efficiency of their inference systems has become in

safetyarxiv-cs-cl

Multi-Channel Feature Fusion and Monte Carlo Dropout for Uncertainty-Aware Diabetic Retinopathy Grading

arXiv:2608.15234v1 Announce Type: cross Abstract: Automated five-stage diabetic retinopathy (DR) grading requires more than high accuracy alone. Medical-grade deployment calls for lesion-aware preproc

Multi-Feature Riemannian Hypergraph for Online Test-Time Adaptation of Motor Imagery Brain-Computer Interface

arXiv:2608.16134v1 Announce Type: new Abstract: In clinical motor imagery brain-computer interface (MI-BCI) decoding, cross-day transferability and online operation remain two critical challenges. Hyp

Multi-Modal Generative Fuzzy System: Fuzzy Inference Guided Large Model Interactive Question Answering Framework

arXiv:2608.14584v1 Announce Type: cross Abstract: In Multimodal Question Answering (MQA), models are required to jointly encode and integrate heterogeneous information from multiple modalities, includ

Near-Optimal Sample Complexity Bounds for Constrained Average-Reward MDPs

arXiv:2509.16586v2 Announce Type: replace Abstract: Recent advances have significantly improved our understanding of the sample complexity of learning in average-reward Markov decision processes (AMDP

Neurosymbolic Embodied Agents

arXiv:2608.16794v1 Announce Type: cross Abstract: Language and vision-language models generate plausible embodied plans but do not guarantee executability, as their outputs can violate environment dyn

nobody takes Altman at his word anymore – nor should they.

nobody takes Altman at his word anymore – nor should they. Altman's claim that OpenAI is slowing their frontier model training in order to focus on safety doesn't pass the sniff test.👃 For several yea

safetygary-marcus--x

NPU Offloading of a Frozen Visual Encoder for Robot Policy Training

arXiv:2608.15002v1 Announce Type: cross Abstract: When a robot policy is trained for a new task or dataset, its visual encoder can be frozen and only its action generation module trained, reducing tra

Observation-Constrained Joint-Space Viewpoint Optimization for Robotic Inspection of Cylindrical Cavities

arXiv:2608.16442v1 Announce Type: new Abstract: Inspection is a core capability in many mobile robotics applications, including industrial facility monitoring, infrastructure maintenance, agriculture,

Offline Reinforcement Learning for Hemodynamic Management of Sepsis in the ICU: a MIMIC-IV Study with Dual Off-Policy Evaluation

arXiv:2608.16482v1 Announce Type: new Abstract: The dosing of intravenous fluids and vasopressors in sepsis is a sequential decision made under uncertainty and guided largely by clinical judgment, whi

One Residual with Three Reuses: A Wristband Front End for Gesture Sensing

arXiv:2608.16542v1 Announce Type: new Abstract: Continuous wrist-worn hand sensing for gesture interfaces and motor symptom monitoring needs an always-on front end that fits inside a coin-cell power b