Search: “safety”

What Does the Caption Really Say? Counterfactual Phrase Intervention for Compositional Data Selection in Vision-Language Pretraining

arXiv:2605.22651v1 Announce Type: new Abstract: CLIP-style contrastive pretraining typically curates web-scale image-text pairs using sample-level filtering signals, often based on pair-level alignmen

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Safety

💯: Why OpenAI keeps taking childish shots at me, is exactly what @FrankRundatz says below: “you have the added audacity of being influentia…

DGX agent

💯: Why OpenAI keeps taking childish shots at me, is exactly what @FrankRundatz says below: “you have the added audacity of being influential enough to move the needle on the timing of OpenAI’s record-

safetygary-marcus--x

22 May 2026

Why Semantic Entropy Fails: Geometry-Aware and Calibrated Uncertainty for Policy Optimization

arXiv:2605.21801v1 Announce Type: cross Abstract: Post-training has become central to improving reasoning and alignment in large language models, where critic-free models enable scalable learning from

'Would You Want an AI Tutor?' Understanding Stakeholder Perceptions of LLM-based Systems in the Classroom

arXiv:2503.02885v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have gained traction in educational settings, often framed as virtual tutors or teaching assistants. Following ea

X-OmniClaw Technical Report: A Unified Mobile Agent for Multimodal Understanding and Interaction

arXiv:2605.05765v2 Announce Type: replace Abstract: Inspired by the development of OpenClaw, there is a growing demand for mobile-based personal agents capable of handling complex and intuitive intera

3D Reconstruction and Knowledge Distillation to Improve Multi-View Image Models to Explore Spike Volume Estimation in Wheat

arXiv:2605.20940v1 Announce Type: new Abstract: Accurate estimation of wheat spike volume is important for yield component analysis and stress resilience assessment, yet field-based measurement remain

A 10,000-Year Global Stochastic Tropical Cyclone Catalog with Wind-Dependent Track Transitions (WHITS)

arXiv:2605.20494v1 Announce Type: new Abstract: Reliable assessment of tropical cyclone (TC) risk is limited by the brevity and spatial sparsity of the historical record, particularly for the rare, hi

A Systematic Comparison between Extractive Self-Explanations and Human Rationales in Text Classification

arXiv:2410.03296v4 Announce Type: replace Abstract: Instruction-tuned LLMs are able to provide extit{an} explanation about their output to users by generating self-explanations, without requiring the

Advantage Collapse in Group Relative Policy Optimization: Diagnosis and Mitigation

arXiv:2605.21125v1 Announce Type: new Abstract: Group Relative Policy Optimization (GRPO), a prominent algorithm within the Reinforcement Learning from Verifiable Rewards (RLVR) framework, has achieve

AFD-INSTRUCTION: A Comprehensive Antibody Instruction Dataset with Functional Annotations for LLM-Based Understanding and Design

arXiv:2602.04916v3 Announce Type: replace-cross Abstract: Large language models (LLMs) have significantly advanced protein representation learning. However, their capacity to interpret and design anti

all these dudes posting about anthropic’s profits without checking the fine print. (see my earlier tweet) there’s some fine print i believe …

all these dudes posting about anthropic’s profits without checking the fine print. (see my earlier tweet) there’s some fine print i believe that n the openai result too, and a lot that has not been di

Always read the fine print: Anthropic is projecting its first (slightly) profitable quarter ever, which is amazing—assuming it actually happ…

Always read the fine print: Anthropic is projecting its first (slightly) profitable quarter ever, which is amazing—assuming it actually happens —but if it does it will be in no small part because they

… and (if am not mistaken) the only one in the top 14 to have never turned a profit. Welcome to the world of vibe investing!

… and (if am not mistaken) the only one in the top 14 to have never turned a profit. Welcome to the world of vibe investing! OpenAI is likely to raise 60 billion in its IPO, more than double Saudi Ara

approximately equals “I am already rich and fuck you if you lose your job to AI”

approximately equals “I am already rich and fuck you if you lose your job to AI” Marc Andreessen to Joe Rogan: Why AI Workers Beat Human Workers⁣ ⁣ 'Never gets drunk. Never gets sick. Never gets depre

Automated Byzantine-Resilient Clustered Decentralized Federated Learning for Battery Intelligence in Connected EVs

arXiv:2605.21115v1 Announce Type: cross Abstract: Federated learning (FL) has emerged as a promising paradigm for managing electric vehicle (EV) battery data in intelligent transportation systems (ITS

AVSD: Adaptive-View Self-Distillation by Balancing Consensus and Teacher-Specific Privileged Signals

arXiv:2605.20643v1 Announce Type: cross Abstract: Self-distillation enables language models to learn on-policy from their own trajectories by using the same model as both student and teacher, with the

Bayesian Preference Learning for Test-Time Steerable Reward Models

arXiv:2602.08819v2 Announce Type: replace-cross Abstract: Reward models are central to aligning language models with human preferences via reinforcement learning (RL). As RL is increasingly applied to

Behavior-Consistent Deep Reinforcement Learning

arXiv:2605.21214v1 Announce Type: new Abstract: Reinforcement learning (RL) often exhibits high variance across training runs, leading to unreliable performance and posing a major challenge to deploym

Beyond Text-to-SQL: An Agentic LLM System for Governed Enterprise Analytics APIs

arXiv:2605.21027v1 Announce Type: new Abstract: Enterprise analytics aims to make organizational data accessible for decision-making, yet non-technical users still face barriers when using traditional

Beyond the Bellman Recursion: A Pontryagin-Guided Framework for Non-Exponential Discounting

arXiv:2605.20996v1 Announce Type: new Abstract: Most value-based and actor--critic reinforcement learning methods rely on Bellman-style recursions, yet these recursions collapse under non-exponential

Biomedical AI may be headed for a replication crisis. (This work below is not about AI-generated reports; it’s about studies of biomedicine …

Biomedical AI may be headed for a replication crisis. (This work below is not about AI-generated reports; it’s about studies of biomedicine that use ML in their methods, and how they are evaluted.) In

Can Microcanonical Langevin Dynamics Leverage Mini-Batch Gradient Noise?

arXiv:2602.06500v2 Announce Type: replace Abstract: Scaling inference methods such as Markov chain Monte Carlo to high-dimensional models remains a central challenge in Bayesian deep learning. A promi

Can Vision Models Truly Forget? Mirage: Representation-Level Certification of Visual Unlearning

arXiv:2605.20282v1 Announce Type: new Abstract: Machine unlearning in Vertical Federated Learning (VFL) has attracted growing interest, yet existing methods certify forgetting solely using output-leve

can’t believe people assume that success on highly verifiable problems in math (where we don’t even know how many tests were performed and h…

can’t believe people assume that success on highly verifiable problems in math (where we don’t even know how many tests were performed and how many might have failed) iautomatically generalize to ever

Checking the math behind the latest headlines from OpenAI and Anthropic, link below:

Gary Marcus examines and fact-checks recent claims made by OpenAI and Anthropic in their public announcements and media coverage. The post links to detailed analysis questioning the mathematical valid

Choose Wisely and Privately: Proactive Client Selection for Fair and Efficient Federated Learning

arXiv:2605.20975v1 Announce Type: new Abstract: Federated Learning enables collaborative model training across decentralized data sources without data transfer. Averaging-based FL is limited by the pr

Comparative Evaluation of Deep Learning Models for Fake Image Detection

arXiv:2605.20971v1 Announce Type: new Abstract: The growing sophistication of GAN-based image manipulation presents significant challenges for digital forensics. This study compares the performance of

Comparing Explanations is Not Enough, Explain the Change: New Standards are Needed to Explain Behavioral Shifts in Large Language Models

arXiv:2602.02304v2 Announce Type: replace-cross Abstract: Large-scale foundation models exhibit behavioral shifts when subjected to interventions such as scaling, fine-tuning, reinforcement learning w

CompilerKV: Risk-Adaptive KV Compression via Offline Experience Compilation

arXiv:2602.08686v2 Announce Type: replace Abstract: Prefill-only KV compression freezes a token subset at the end of prefill and decodes from it without further eviction. The retention decision is the

Complementing reinforcement learning with SFT through logit averaging in the post training of LLMs

arXiv:2605.20555v1 Announce Type: new Abstract: We introduce a novel method that averages the logits of a frozen reference policy (e.g., SFT) and a trainable policy, and incorporate the method into Gr

Conditional Equivalence of DPO and RLHF: Implicit Assumption, Failure Modes, and Provable Alignment

arXiv:2605.20834v1 Announce Type: cross Abstract: Direct Preference Optimization (DPO) has emerged as a popular alternative to Reinforcement Learning from Human Feedback (RLHF), offering theoretical e

Consistently Informative Soft-Label Temperature for Knowledge Distillation

arXiv:2605.20357v1 Announce Type: new Abstract: Knowledge distillation (KD) transfers knowledge from a high-capacity teacher to a compact student by matching their predictive distributions, with tempe

Correcting Stochastic Update Bias in Preconditioned Language Model Optimizers

arXiv:2605.20756v1 Announce Type: new Abstract: Preconditioned optimizers are central to language model training, but their stochastic update rules are usually treated as direct approximations to popu

CRAFT: Conflict-Resolved Aggregation for Federated Training

arXiv:2605.21317v1 Announce Type: new Abstract: The aggregation of conflicting client updates remains a fundamental bottleneck in federated learning (FL) under heterogeneous data distributions. Naive

Cross-lingual robustness of LLM-brain alignment and its computational roots

arXiv:2605.21049v1 Announce Type: new Abstract: Large language models (LLMs) reliably predict neural activity during language comprehension and transformer depth has been interpreted as mirroring hier

Cumulative Meta-Learning from Active Learning Queries for Robustness to Spurious Correlations

arXiv:2605.20771v1 Announce Type: new Abstract: Spurious correlations in real-world datasets cause machine learning models to rely on irrelevant patterns, undermining reliability, generalization, and

Data-Efficient Neural Operator Training via Physics-Based Active Learning

arXiv:2605.21348v1 Announce Type: new Abstract: Solving partial differential equations with neural operators significantly reduces computational costs but remains bottlenecked by high training data re

Decision-Path Patterns as Tree Reliability Signals: Path-based Adaptive Weighting for Random Forest Classification

arXiv:2605.20716v1 Announce Type: new Abstract: Random forests aggregate tree votes by simple majority, treating all trees as equally informative. We observe that the topological pattern along each tr

Decomposing MXFP4 quantization error for LLM reinforcement learning: reducible bias, recoverable deadzone, and an irreducible floor

arXiv:2605.20402v1 Announce Type: new Abstract: MXFP4 arithmetic can dramatically accelerate reinforcement learning (RL) post-training of large language models (LLMs), yet the quantization error intro

DeCoR: Design and Control Co-Optimization for Urban Streets Using Reinforcement Learning

arXiv:2605.21311v1 Announce Type: new Abstract: Modern vision systems can detect, track, and forecast urban actors at scale, yet translating perception outputs to urban design remains limited. We intr

Decoupling Communication from Policy: Robust MARL under Bandwidth Constraints

arXiv:2605.21085v1 Announce Type: cross Abstract: Communication enables coordination in multi-agent reinforcement learning (MARL), but many real-world applications, e.g., search-and-rescue with drone

Deep Attention Reweighting: Post-Hoc Attention-Based Feature Aggregation in CNNs for Disentangling Core and Spurious Features under Spurious Correlations

arXiv:2605.20732v1 Announce Type: new Abstract: Convolutional Neural Networks (CNNs) often exploit spurious correlations in datasets, learning superficially predictive yet causally irrelevant features

DelTA: Discriminative Token Credit Assignment for Reinforcement Learning from Verifiable Rewards

arXiv:2605.21467v1 Announce Type: cross Abstract: Reinforcement learning from verifiable rewards (RLVR) has emerged as a central technique for improving the reasoning capabilities of large language mo

Design for Manufacturing: A Manufacturability Knowledge-Integrated Reinforcement Learning Framework for Free-Form Pipe Routing in Aeroengines

arXiv:2605.20644v1 Announce Type: new Abstract: Design for manufacturing plays a critical role in advanced aeroengine development, where complex components necessitate careful consideration of manufac

Disentangling Bias by Modeling Intra- and Inter-modal Causal Attention for Multimodal Sentiment Analysis

arXiv:2508.04999v2 Announce Type: replace Abstract: Multimodal sentiment analysis (MSA) aims to understand human emotions by integrating information from multiple modalities, such as text, audio, and

Distributed Direct Preference Optimization

arXiv:2605.20696v1 Announce Type: new Abstract: Preference-based reinforcement learning (RL) is a key paradigm for aligning policies with human judgments, yet its theoretical behavior in distributed s

Distribution-Aware Reward: Reinforcement Learning over Predictive Distributions for LLM Regression

arXiv:2605.20740v1 Announce Type: cross Abstract: Large language models can predict real-valued quantities from heterogeneous inputs such as text, code, and molecular strings, but most training object

Distributional Alignment as a Criterion for Designing Task Vectors in In-Context Learning

arXiv:2605.20730v1 Announce Type: new Abstract: In-context learning (ICL) allows large language models (LLMs) to adapt to new tasks through demonstrations, yet it suffers from escalating inference cos