HN dakheera47 2/13/2026

Show HN: Fuzzy-matching messy job board data against the UK Gov Visa Registry

Fuzzy-matching tool for matching job board data against UK visa sponsor registry. Data processing application with practical developer utility.

HN thomaskiko 2/13/2026

Show HN: AI-powered video creation web app

Web app for AI-powered video generation with intuitive UI. LLM/generative AI application but limited developer tooling focus.

HN helloplanets 2/13/2026

Prompts and responses from DeepMind's Aletheia on research level math problems

Collection of prompts and responses from DeepMind's Aletheia model on advanced mathematics problems.

HN demirdev 2/13/2026

Show HN: ToMusic – Turn text prompts into original songs with AI (iOS)

iOS app converting text prompts to complete songs using AI models. LLM application but music-focused, limited developer relevance.

DT Helder Burato Berto 2/13/2026

Building Shareable AI Agent Skills

Short article discussing building reusable AI agent skills to avoid repetitive instruction setup across sessions with Claude Code.

BL 2/13/2026

GPT-5.2 derives a new result in theoretical physics

GPT-5.2 proposes new gluon amplitude formula in theoretical physics, later formally proved by OpenAI and academic collaborators.

BL 2/13/2026

Introducing Lockdown Mode and Elevated Risk labels in ChatGPT

ChatGPT adds Lockdown Mode and Elevated Risk labels to defend against prompt injection and AI-driven data exfiltration.

BL 2/13/2026

Beyond rate limits: scaling access to Codex and Sora

OpenAI describes real-time access system combining rate limits and credits to power continuous access to Sora and Codex.

BL 2/13/2026

Scaling social science research

GABRIEL is an open-source toolkit using GPT to convert qualitative text and images into quantitative data for social science research.

HF Baorong Shi, Bo Cui, Boyuan Jiang, Deli Yu, Fang Qian, Haihua Yang, Huichao Wang, Jiale Chen, Jianfei Pan, Jieqiong Cao, Jinghao Lin, Kai Wu, Lin Yang, Shengsheng Yao, Tao Chen, Xiaojun Xiao, Xiaozhong Ji, Xu Wang, Yijun He, Zhixiong Yang 2/13/2026

MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs

MedXIAOHE: medical vision-language foundation model with entity-aware pretraining for clinical applications, achieves SOTA on medical benchmarks.

HN AkshatRaj00 2/13/2026

Small Language Models (SLMs) vs. Large Language Models (LLMs)

Research comparing small language models vs large language models, focusing on task-optimized, on-device variants with reduced computational requirements.

HF Rui Cai, Jun Guo, Xinze He, Piaopiao Jin, Jie Li, Bingxuan Lin, Futeng Liu, Wei Liu, Fei Ma, Kun Ma, Feng Qiu, Heng Qu, Yifei Su, Qiao Sun, Dong Wang, Donghao Wang, Yunhong Wang, Rujie Wu, Diyun Xiang, Yu Yang, Hangjun Ye, Yuan Zhang, Quanyun Zhou 2/13/2026

Xiaomi-Robotics-0: An Open-Sourced Vision-Language-Action Model with Real-Time Execution

Xiaomi-Robotics-0: open-sourced vision-language-action model for robotics with cross-embodiment pretraining and optimized real-time execution.

HN xuyanmei 2/13/2026

Show HN: Seedance AShow HN: Seedance AI Video Generation (Next.js, Drizzle)

Seedance: Next.js-based tool generating videos and images from text prompts using AI with template and credit-based pricing.

HF Liangzhi Shi, Shuaihang Chen, Feng Gao, Yinuo Chen, Kang Chen, Tonghe Zhang, Hongzhi Zhang, Weinan Zhang, Chao Yu, Yu Wang 2/13/2026

RLinf-Co: Reinforcement Learning-Based Sim-Real Co-Training for VLA Models

RL-based sim-real co-training method for vision-language-action models that combines simulation and real-world robot data using reinforcement learning instead of supervised fine-tuning.

HN shabers 2/13/2026

Show HN: Codeman – a blunt launcher forcing you to pick a Codex permission level

Codeman: open source security-focused launcher for Codex with permission levels, session resumption, and webhook notifications.

Ax Natalia Abarca, Andr\'es Carvallo, Claudia L\'opez Moncada, Felipe Bravo-Marquez 2/13/2026

Explaining AI Without Code: A User Study on Explainable AI

User study examining explainable AI methods in no-code ML platforms, addressing transparency gaps for non-technical users in sensitive domains.

Ax Zituo Chen, Haixu Wu, Sili Deng 2/13/2026

Latent Generative Solvers for Generalizable Long-Term Physics Simulation

Latent Generative Solvers framework using VAE and Transformer for long-horizon physics simulation across heterogeneous PDE systems with uncertainty mechanisms.

Ax John M. Willis 2/13/2026

The PBSAI Governance Ecosystem: A Multi-Agent AI Reference Architecture for Securing Enterprise AI Estates

Reference architecture for securing enterprise AI estates with multi-agent governance covering LLMs, RAG pipelines, agents, and shared infrastructure security.

Ax Bo Pan, Xuan Kan, Kaitai Zhang, Yan Yan, Shunwen Tan, Zihao He, Zixin Ding, Junjie Wu, Liang Zhao 2/13/2026

Bi-Level Prompt Optimization for Multimodal LLM-as-a-Judge

Bi-level prompt optimization method aligning LLM-based evaluations with human judgments without costly fine-tuning, applicable across tasks and datasets.

Ax Ruipeng Wang, Yuxin Chen, Yukai Wang, Chang Wu, Junfeng Fang, Xiaodong Cai, Qi Gu, Hui Su, An Zhang, Xiang Wang, Xunliang Cai, Tat-Seng Chua 2/13/2026

AgentNoiseBench: Benchmarking Robustness of Tool-Using LLM Agents Under Noisy Condition

AgentNoiseBench benchmarks robustness of tool-using LLM agents under realistic noisy conditions versus idealized settings to identify deployment gaps.

Ax Yihang Yao, Zhepeng Cen, Haohong Lin, Shiqi Liu, Zuxin Liu, Jiacheng Zhu, Zhang-Wei Hong, Laixi Shi, Ding Zhao 2/13/2026

Pushing Forward Pareto Frontiers of Proactive Agents with Behavioral Agentic Optimization

Agentic RL approach optimizing proactive LLM agents' Pareto frontiers through behavioral optimization for efficient multi-turn task completion.

Ax Bang Nguyen, Dominik So\'os, Qian Ma, Rochana R. Obadage, Zack Ranjan, Sai Koneru, Timothy M. Errington, Shakhlo Nematova, Sarah Rajtmajer, Jian Wu, Meng Jiang 2/13/2026

ReplicatorBench: Benchmarking LLM Agents for Replicability in Social and Behavioral Sciences

ReplicatorBench evaluates LLM agents on replicability of social/behavioral science papers, addressing data availability variations and reproduction vs replication gaps.

Ax Heejeong Nam, Quentin Le Lidec, Lucas Maes, Yann LeCun, Randall Balestriero 2/13/2026

Causal-JEPA: Learning World Models through Object-Level Latent Interventions

Object-centric world model extending masked joint embedding prediction with object-level latent interventions for capturing interaction-dependent dynamics.

Ax Sina Tayebati, Divake Kumar, Nastaran Darabi, Davide Ettori, Ranganath Krishnan, Amit Ranjan Trivedi 2/13/2026

TRACER: Trajectory Risk Aggregation for Critical Episodes in Agentic Reasoning

TRACER framework for estimating uncertainty in multi-turn agentic interactions by detecting trajectory-level breakdowns like looping and tool misuse.

Ax Chengrui Qu, Christopher Yeh, Kishan Panaganti, Eric Mazumdar, Adam Wierman 2/13/2026

Distributionally Robust Cooperative Multi-Agent Reinforcement Learning via Robust Value Factorization

Value-factorization method for cooperative multi-agent RL addressing sim-to-real gaps and environmental uncertainties while maintaining decentralized execution.

Ax Zhengbo Jiao, Shaobo Wang, Zifan Zhang, Wei Wang, Bing Zhao, Hu Wei, Linfeng Zhang 2/13/2026

Credit Where It is Due: Cross-Modality Connectivity Drives Precise Reinforcement Learning for MLLM Reasoning

Analysis of visual-textual coupling in multimodal RL for MLLMs showing only ~15% of tokens exhibit strong cross-modal connectivity during reasoning.

Ax Faouzi El Yagoubi, Ranwa Al Mallah, Godwin Badu-Marfo 2/13/2026

AgentLeak: A Full-Stack Benchmark for Privacy Leakage in Multi-Agent LLM Systems

First full-stack benchmark measuring privacy leakage in multi-agent LLM systems across inter-agent messages, shared memory, and tool arguments covering 1,000 scenarios.

Ax Hong Su 2/13/2026

Human-Inspired Continuous Learning of Internal Reasoning Processes: Learning How to Think for Adaptive AI Systems

Framework for continuous learning of internal reasoning processes and action scheduling policies in AI systems for sustained adaptation in dynamic environments.

Ax Jiawei Zhu, Wei Chen, Ruichu Cai 2/13/2026

CausalAgent: A Conversational Multi-Agent System for End-to-End Causal Inference

Multi-agent conversational system automating causal inference workflows by lowering technical barriers and handling algorithm selection, data quality, and result interpretation.

Ax Hanbing Liu, Chunhao Tian, Nan An, Ziyuan Wang, Pinyan Lu, Changyuan Yu, Qi Qi 2/13/2026

Budget-Constrained Agentic Large Language Models: Intention-Based Planning for Costly Tool Use

Framework for tool-augmented LLM agents operating under monetary budget constraints using intention-based planning to handle expensive tool invocations.

Ax Aditya Taparia, Som Sagar, Ransalu Senanayake 2/13/2026

Learning to Configure Agentic AI Systems

Method to automatically configure LLM-based agents by learning query-wise decisions for workflows, tools, token budgets, and prompts instead of using fixed templates.

Ax Jingdi Chen, Hanqing Yang, Zongjun Liu, Carlee Joe-Wong 2/13/2026

The Five Ws of Multi-Agent Communication: Who Talks to Whom, When, What, and Why -- A Survey from MARL to Emergent Language and LLMs

Survey of multi-agent communication in MARL, robotics, and LLM systems covering who communicates, what/when/why communication occurs, and emergent language.

Ax Nikhil Verma, Minjung Kim, JooYoung Yoo, Kyung-Min Jin, Manasa Bharadwaj, Kevin Ferreira, Ko Keun Kim, Youngjoon Kim 2/13/2026

MAPLE: Modality-Aware Post-training and Learning Ecosystem

MAPLE improves RL post-training for multimodal language models by weighting modalities based on task relevance, reducing variance and improving robustness to missing/added signals.

Ax Yiming Gao, Zhen Wang, Jefferson Chen, Mark Antkowiak, Mengzhou Hu, JungHo Kong, Dexter Pratt, Jieyuan Liu, Enze Ma, Zhiting Hu, Eric P. Xing 2/13/2026

scPilot: Large Language Model Reasoning Toward Automated Single-Cell Analysis and Discovery

scPilot framework enables LLMs to perform omics-native reasoning on single-cell RNA-seq data, automating cell-type annotation and trajectory reconstruction through tool integration.

Ax Aman Mehta 2/13/2026

When Agents Disagree With Themselves: Measuring Behavioral Consistency in LLM-Based Agents

Studies behavioral consistency of LLM-based agents across 3,000 runs, finding 2.0-4.2 distinct action sequences per 10 runs and showing variance predicts task failure.

Ax Shuo Lu, Jianjie Cheng, Yinuo Xu, Yongcan Yu, Lijun Sheng, Peijie Wang, Siru Jiang, Yongguan Hu, Run Ling, Yihua Shao, Ao Ma, Wei Feng, Lingxiao He, Meng Wang, Qianlong Xie, Xingxing Wang, Ran He, Jian Liang 2/13/2026

Do MLLMs Really Understand Space? A Mathematical Reasoning Evaluation

Evaluates multimodal LLMs on mathematical spatial reasoning tasks, finding significant gaps (60% vs 95% human accuracy) in 2D/3D relation parsing and manipulation.

Ax Tianxiang Xu, Jiayi Liu, Yixuan Tong, Jialu Xu, Yunqing Wei, Kaiwen Feng, PanPan Hou, Kangping Yin, Jiyuan Hu, Hao Zhou, Zhenxin Ma, Jian Xu, Guanjun Jiang 2/13/2026

Quark Medical Alignment: A Holistic Multi-Dimensional Alignment and Collaborative Optimization Paradigm

Multi-dimensional alignment framework for medical LLMs addressing RLHF limitations in high-stakes applications through verifiable rewards and collaborative optimization.

Ax E Fan, Lisong Shi, Zhengtong Li, Chih-yung Wen 2/13/2026

PhyNiKCE: A Neurosymbolic Agentic Framework for Autonomous Computational Fluid Dynamics

PhyNiKCE combines neurosymbolic reasoning with agentic LLMs for computational fluid dynamics, addressing probabilistic LLM limitations through symbolic constraints and physics-aware planning.

Ax Longyuan Zhu, Hairan Hua, Linlin Miao, Bing Zhao 2/13/2026

Benchmark Health Index: A Systematic Framework for Benchmarking the Benchmarks of LLMs

Benchmark Health Index framework audits LLM evaluation sets to detect score inflation and selective reporting, providing systematic quality assessment of benchmark reliability.

Ax Edward Y. Chang 2/13/2026

Right for the Wrong Reasons: Epistemic Regret Minimization for Causal Rung Collapse in LLMs

Identifies causal origin of LLM shortcuts through Rung Collapse theory, where autoregressive training fails to distinguish association from intervention, proposing epistemic regret minimization.

Ax Chengwei Ma, Zhen Tian, Zhou Zhou, Zhixian Xu, Xiaowei Zhu, Xia Hua, Si Shi, F. Richard Yu 2/13/2026

Beyond Pixels: Vector-to-Graph Transformation for Reliable Schematic Auditing

Vector-to-Graph pipeline converts CAD diagrams to structured graphs to improve multimodal LLM understanding of engineering schematics and topology, addressing pixel-driven limitations.

Ax Xin Xu, Tong Yu, Xiang Chen, Haoliang Wang, Julian McAuley, Saayan Mitra 2/13/2026

ThinkRouter: Efficient Reasoning via Routing Thinking between Latent and Discrete Spaces

ThinkRouter improves reasoning efficiency by routing thinking between latent and discrete reasoning spaces based on model confidence dynamics analysis.

Ax Weihong Lin, Lin Sun, Qilong Shi, Aomufei Yuan, Yuxuan Tian, Zhengyang Wang, Guangxiang Zhao, Xiangzheng Zhang, Tong Yang 2/13/2026

Beyond Parameter Arithmetic: Sparse Complementary Fusion for Distribution-Aware Model Merging

Sparse complementary fusion method for distribution-aware LLM merging that reduces parameter interference and improves generalization over existing weight-space heuristics.

Ax Thomas Jiralerspong, Trenton Bricken 2/13/2026

Cross-Architecture Model Diffing with Crosscoders: Unsupervised Discovery of Differences Between LLMs

Crosscoders enable unsupervised cross-architecture model diffing to identify differences in LLM internal representations and uncover safety-critical behaviors.

Ax Songlin Lyu, Lujie Ban, Zihang Wu, Tianqi Luo, Jirong Liu, Chenhao Ma, Yuyu Luo, Nan Tang, Shipeng Qi, Heng Lin, Yongchao Liu, Chuntao Hong 2/13/2026

Text2GQL-Bench: A Text to Graph Query Language Benchmark [Experiment, Analysis & Benchmark]

Text2GQL-Bench benchmark for evaluating LLM systems translating natural language into graph query language for graph database analysis and manipulation.

Ax Zibo Xiao, Jun Sun, Junjie Chen 2/13/2026

AIR: Improving Agent Safety through Incident Response

AIR: First incident response framework for LLM agent systems defining protocols for responding to, containing, and recovering from agent failures after deployment.

Ax Aladin Djuhera, Swanand Ravindra Kadhe, Farhan Ahmed, Holger Boche 2/13/2026

TSR: Trajectory-Search Rollouts for Multi-Turn RL of LLM Agents

TSR (Trajectory-Search Rollouts) training approach for multi-turn RL of LLM agents that improves exploration and avoids mode collapse with sparse/delayed rewards.

Ax Jinfang Wang, Jiajie Liu, Jianwei Wu, Ziqin Luo, Zhen Chen, Chunlei Li, Biao Han, Tao Deng, Yi Li, Shuanglong Li, Lin Liu 2/13/2026

RELATE: A Reinforcement Learning-Enhanced LLM Framework for Advertising Text Generation

RL-enhanced LLM framework for advertising text generation that jointly optimizes for both generation quality and online performance metrics like CTR.

Ax Yihao Liu, Ziyun Zhang, Zile He, Huaqian Cai 2/13/2026

FlowMind: Execute-Summarize for Structured Workflow Generation from LLM Reasoning

FlowMind framework translates LLM reasoning and tool use into structured workflows using execute-summarize mechanism to avoid interference between task solving and workflow generation.

Ax Lingyong Yan, Jiulong Wu, Dong Xie, Weixian Shi, Deguo Xia, Jizhou Huang 2/13/2026

Beyond End-to-End Video Models: An LLM-Based Multi-Agent System for Educational Video Generation

LAVES: LLM-based multi-agent system for generating instructional videos from educational problems with hierarchical decomposition and strict logical rigor.