HN oger 2/19/2026

HumanCompiler – Compile humans into AI agents – a Claude Code plugin

Claude Code plugin that converts human behavior into installable AI agent plugins through interviews and artifact analysis.

HN aarvinroshin 2/19/2026

AI Agents Are Taking America by Storm

Overview of AI agents becoming mainstream, discussing tools like Claude Code and their ability to automate weeks of work into hours.

HN petruspennanen 2/19/2026

AgentPuzzles – API‑first timed puzzle arena and public leaderboard for AI agents

API-first puzzle arena and leaderboard for testing AI agents across CAPTCHAs, geolocation, lateral thinking, and research problems.

Ax Jieyi Bi, Zhiguang Cao, Jianan Zhou, Wen Song, Yaoxin Wu, Jie Zhang, Yining Ma, Cathy Wu 2/19/2026

Towards Efficient Constraint Handling in Neural Solvers for Routing Problems

Construct-and-Refine (CaR) method for efficient constraint handling in neural solvers for complex routing problems with hard constraints.

Ax Cameron Cagan, Pedram Fard, Jiazi Tian, Jingya Cheng, Shawn N. Murphy, Hossein Estiri 2/19/2026

Optimization Instability in Autonomous Agentic Workflows for Clinical Symptom Detection

Investigates optimization instability in autonomous agentic workflows using Pythia framework, characterizing failure modes in clinical symptom detection systems.

Ax Hang Li, Kaiqi Yang, Xianxuan Long, Fedor Filippov, Yucheng Chu, Yasemin Copur-Gencturk, Peng He, Cory Miller, Namsoo Shin, Joseph Krajcik, Hui Liu, Jiliang Tang 2/19/2026

How Uncertain Is the Grade? A Benchmark of Uncertainty Metrics for LLM-Based Automatic Assessment

Benchmark measuring uncertainty metrics for LLM-based automatic assessment in education, addressing probabilistic output reliability challenges.

Ax Amir Hosseinian, MohammadReza Zare Shahneh, Umer Mansoor, Gilbert Szeto, Kirill Karlin, Nima Aghaeepour 2/19/2026

Evidence-Grounded Subspecialty Reasoning: Evaluating a Curated Clinical Intelligence Layer on the 2025 Endocrinology Board-Style Examination

Evaluates Mirror, an evidence-grounded clinical reasoning system, against GPT-5 variants on endocrinology board exams with nuanced subspecialty knowledge.

Ax Martin Klissarov, Jonathan Cook, Diego Antognini, Hao Sun, Jingling Li, Natasha Jaques, Claudiu Musat, Edward Grefenstette 2/19/2026

Improving Interactive In-Context Learning from Natural Language Feedback

Framework for improving LLMs' ability to adapt dynamically using interactive natural language feedback in collaborative settings, beyond static training paradigms.

Ax Thinh Hung Truong, Jey Han Lau, Jianzhong Qi 2/19/2026

GPSBench: Do Large Language Models Understand GPS Coordinates?

GPSBench benchmark evaluates LLM geospatial reasoning with 57,800 samples across 17 tasks for navigation, robotics, and mapping applications.

Ax Kaiqu Liang, Julia Kruk, Shengyi Qian, Xianjun Yang, Shengjie Bi, Yuanshun Yao, Shaoliang Nie, Mingyang Zhang, Lijuan Liu, Jaime Fern\'andez Fisac, Shuyan Zhou, Saghar Hosseini 2/19/2026

Learning Personalized Agents from Human Feedback

PAHF framework enables AI agents to learn and adapt to individual user preferences dynamically through human feedback, addressing personalization and preference drift.

Ax Sushant Mehta, Logan Ritchie, Suhaas Garre, Nick Heiner, Edwin Chen 2/19/2026

EnterpriseGym Corecraft: Training Generalizable Agents on High-Fidelity RL Environments

EnterpriseGym introduces CoreCraft, a high-fidelity RL environment with 2,500+ entities for training generalizable AI agents on enterprise customer support simulations.

Ax Hiroaki Yamanaka, Daisuke Miyashita, Takashi Toi, Asuka Maki, Taiga Ikeda, Jun Deguchi 2/19/2026

Revolutionizing Long-Term Memory in AI: New Horizons with High-Capacity and High-Speed Storage

Design concepts for memory systems essential for artificial superintelligence, exploring high-capacity storage approaches.

Ax Yun-Shiuan Chuang, Chaitanya Kulkarni, Alec Chiu, Avinash Thangali, Zijie Pan, Shivani Shekhar, Yirou Ge, Yixi Li, Uma Kona, Linsey Pang, Prakhar Mehrotra 2/19/2026

Toward Scalable Verifiable Reward: Proxy State-Based Evaluation for Multi-turn Tool-Calling LLM Agents

Proxy State-Based Evaluation method for benchmarking multi-turn tool-calling LLM agents with LLM-driven simulation backends.

Ax Marissa A. Weis, Maciej Wo{\l}czyk, Rajai Nasser, Rif A. Saurous, Blaise Ag\"uera y Arcas, Jo\~ao Sacramento, Alexander Meulemans 2/19/2026

Multi-agent cooperation through in-context co-player inference

In-context co-player inference enabling multi-agent cooperation through learning-aware agents in reinforcement learning.

Ax Philipp Schoenegger, Matt Carlson, Chris Schneider, Chris Daly 2/19/2026

Verifiable Semantics for Agent-to-Agent Communication

Certification protocol for multi-agent communication ensuring semantic consistency between agents using stimulus-meaning model.

Ax Arun Vignesh Malarkkan, Wangyang Ying, Yanjie Fu 2/19/2026

Causally-Guided Automated Feature Engineering with Multi-Agent Reinforcement Learning

CAFE framework combining causal discovery with multi-agent RL for automated feature engineering on tabular data.

Ax Zihao Li, Fabrizio Russo 2/19/2026

Leveraging Large Language Models for Causal Discovery: a Constraint-based, Argumentation-driven Approach

Causal discovery using LLMs with constraint-based argumentation framework to uncover causal relations from data.

Ax Felix Fricke, Simon Malberg, Georg Groh 2/19/2026

Framework of Thoughts: A Foundation Framework for Dynamic and Optimized Reasoning based on Chains, Trees, and Graphs

Framework of Thoughts unifying Chain, Tree, and Graph of Thought prompting with dynamic adaptation and hyperparameter optimization.

Ax Vered Tohar, Tsahi Hayat, Amir Leshem 2/19/2026

Creating a digital poet

Seven-month poetry workshop shaping LLM into digital poet through iterative in-context expert feedback without retraining.

Ax Yangjie Xu, Lujun Li, Lama Sleem, Niccolo Gentile, Yewei Song, Yiqun Wang, Siming Ji, Wenbo Wu, Radu State 2/19/2026

Agent Skill Framework: Perspectives on the Potential of Small Language Models in Industrial Environments

Investigation of Agent Skill framework effectiveness with small language models in industrial environments for context engineering and hallucination reduction.

Ax Stephan Rabanser, Sayash Kapoor, Peter Kirgis, Kangheng Liu, Saiteja Utpala, Arvind Narayanan 2/19/2026

Towards a Science of AI Agent Reliability

Framework for evaluating AI agent reliability beyond accuracy metrics, examining consistency, perturbation robustness, and operational flaws.

Ax Weiwen Su, Yuhan Zhou, Zihan Wang, Naoki Yoshinaga, Masashi Toyoda 2/19/2026

What Persona Are We Missing? Identifying Unknown Relevant Personas for Faithful User Simulation

PICQ dataset and method for identifying unknown relevant personas in user simulation for dialogue systems.

Ax Mengyun Liu, Shanshan Huang, Jianan Jiang 2/19/2026

EdgeNav-QE: QLoRA Quantization and Dynamic Early Exit for LAM-based Navigation on Edge Devices

EdgeNav-QE framework combining QLoRA quantization and dynamic early-exit for deploying Large Action Models on edge devices.

Ax Warren Johnson 2/19/2026

The Perplexity Paradox: Why Code Compresses Better Than Math in LLM Prompts

Validation of code compression tolerance in LLM prompts across six benchmarks, explaining why code compresses better than mathematical reasoning.

Ax Inwon Kang, Parikshit Ram, Yi Zhou, Horst Samulowitz, Oshani Seneviratne 2/19/2026

Language Model Representations for Efficient Few-Shot Tabular Classification

LLM representations for few-shot tabular classification on heterogeneous web data structures.

Ax Pranav Bhandari, Usman Naseem, Mehwish Nasim 2/19/2026

Do Personality Traits Interfere? Geometric Limitations of Steering in Large Language Models

Analysis of geometric relationships between Big Five personality steering vectors in LLaMA and Mistral models.

Ax Andrius Mat\v{s}enas, Anet Lello, T\~onis Lees, Hans Peep, Kim Lilii Tamm 2/19/2026

Can LLMs Assess Personality? Validating Conversational AI for Trait Profiling

Validation study comparing Big Five personality scores from LLM conversations against IPIP-50 questionnaire with moderate convergent validity.

Ax Karun Sharma, Vidushee Vats, Shengzhi Li, Yuxiang Wang, Zhongtian Sun, Prayag Tiwari 2/19/2026

Preference Optimization for Review Question Generation Improves Writing Quality

IntelliReward model using preference optimization to improve review question generation quality beyond surface-level queries.

Ax David Y. Liu, Aditya Joshi, Paul Dawson 2/19/2026

Narrative Theory-Driven LLM Methods for Automatic Story Generation and Understanding: A Survey

Survey of narrative theory-driven LLM methods for story generation and understanding, proposing taxonomy of narrative datasets and theories.

Ax Ha Na Cho, Sairam Sutari, Alexander Lopez, Hansen Bow, Kai Zheng 2/19/2026

Building Safe and Deployable Clinical Natural Language Processing under Temporal Leakage Constraints

Clinical NLP models addressing temporal and lexical leakage in hospital discharge planning to prevent overconfident predictions.

Ax Md Asiful Islam, Mihai Surdeanu 2/19/2026

A Lightweight Explainable Guardrail for Prompt Safety

Multi-task learning method for classifying unsafe prompts with explainability, trained on synthetic data to mitigate LLM confirmation biases.

Ax Jingyi Xu, Xingyu Ren, Zhiqiang You, Yumeng Zhang, Zhoupeng Shou 2/19/2026

Decoupling Strategy and Execution in Task-Focused Dialogue via Goal-Oriented Preference Optimization

GOPO: hierarchical RL framework for task-oriented dialogue decoupling strategy planning from response generation using preference optimization.

Ax Yunhao Liu, Zian Jia, Xinyu Gao, Kanjun Xu, Yun Xiong 2/19/2026

Rethinking Soft Compression in Retrieval-Augmented Generation: A Query-Conditioned Selector Perspective

Query-conditioned soft compression approach for RAG reducing context length and redundancy while maintaining performance.

Ax Annie Wong, Aske Plaat, Thomas B\"ack, Niki van Stein, Anna V. Kononova 2/19/2026

State Design Matters: How Representations Shape Dynamic Reasoning in Large Language Models

Systematic study of how state representation (granularity, structure, language) affects LLM dynamic reasoning in interactive environments.

Ax Jinxiang Xie, Zihao Li, Wei He, Rui Ding, Shi Han, Dongmei Zhang 2/19/2026

CAST: Achieving Stable LLM-based Text Analysis for Data Analytics

CAST: consistency framework for stable LLM-based text analysis on tabular data via algorithmic prompting and token selection.

Ax Guoshan Liu, Bin Zhu, Yian Li, Jingjing Chen, Chong-Wah Ngo, Yu-Gang Jiang 2/19/2026

Enhancing Action and Ingredient Modeling for Semantically Grounded Recipe Generation

Semantically grounded recipe generation from food images using MLLMs with action/ingredient validation framework.

Ax Daehoon Gwak, Minseo Jung, Junwoo Park, Minho Park, ChaeHun Park, Junha Hyung, Jaegul Choo 2/19/2026

Not the Example, but the Process: How Self-Generated Examples Enhance LLM Reasoning

Study shows LLM reasoning improves via self-generated examples through process enhancement rather than example quality.

Ax Most. Sharmin Sultana Samu, Nafisa Khan, Kazi Toufique Elahi, Tasnuva Binte Rahman, Md. Rakibul Islam, Farig Sadeque 2/19/2026

AI as Teammate or Tool? A Review of Human-AI Interaction in Decision Support

Reviews human-AI interaction literature distinguishing AI as tool vs. teammate across interaction design, trust, and healthcare.

Ax Dhiman Goswami, Jai Kruthunz Naveen Kumar, Sanchari Das 2/19/2026

NLP Privacy Risk Identification in Social Media (NLP-PRISM): A Survey

NLP-PRISM: framework surveying privacy risks from NLP processing of social media containing PII and behavioral metadata.

Ax Berry Gerrits 2/19/2026

Playing With AI: How Do State-Of-The-Art Large Language Models Perform in the 1977 Text-Based Adventure Game Zork?

Evaluates problem-solving and reasoning capabilities of contemporary LLMs through performance in Zork text adventure game.

Ax Chuyang Ye, Haoxian Jing, Qinting Jiang, Yixi Lin, Qiang Li, Xing Tang, Jingyan Jiang 2/19/2026

Test-Time Adaptation for Tactile-Vision-Language Models

Test-time adaptation methods for tactile-vision-language models handling modality-wise reliability under distribution shifts.

Ax Zhenxing Xu, Brikit Lu, Weidong Bao, Zhengqiu Zhu, Junsong Zhang, Hui Yan, Wenhao Lu, Ji Wang 2/19/2026

Fly0: Decoupling Semantic Grounding from Geometric Planning for Zero-Shot Aerial Navigation

Fly0 framework decouples semantic reasoning (MLLM) from geometric planning for zero-shot aerial navigation tasks.

Ax Jingjing Fan, Yushan Liu, Shoujie Li, Botao Ren, Siyuan Li, Xiao-Ping Zhang, Wenbo Ding, Zhidong Deng 2/19/2026

FUTURE-VLA: Forecasting Unified Trajectories Under Real-time Execution

FUTURE-VLA: vision-language architecture for real-time robotic control combining long-horizon trajectory forecasting with reduced latency.

Ax Paul Tschisgale, Peter Wulff 2/19/2026

Evidence for Daily and Weekly Periodic Variability in GPT-4o Performance

Study reveals GPT-4o exhibits daily and weekly performance variability, threatening reproducibility of LLM-based research.

Ax Maijunxian Wang, Yijiang Li, Bingyang Wang, Tianwei Zhao, Ran Ji, Qingying Gao, Emmy Liu, Hokin Deng, Dezhi Luo 2/19/2026

Egocentric Bias in Vision-Language Models

FlipSet benchmark evaluates visual perspective taking in 103 vision-language models through 180-degree character rotation tasks.

Ax Pengcheng Zhou, Haochen Li, Zhiqiang Nie, JiaLe Chen, Qing Gong, Weizhen Zhang, Chun Yu 2/19/2026

Understand Then Memory: A Cognitive Gist-Driven RAG Framework with Global Semantic Diffusion

CogitoRAG: RAG framework inspired by human episodic memory to improve semantic integrity and reduce retrieval deviations in LLM-based systems.

Ax Rujikorn Charakorn, Edoardo Cetin, Shinnosuke Uesaka, Robert Tjarko Lange 2/19/2026

Doc-to-LoRA: Learning to Instantly Internalize Contexts

Doc-to-LoRA method distills long document contexts into LoRA parameters to reduce transformer attention cost in LLM inference.

Ax Pengfei Zhang, Tianxin Xie, Minghao Yang, Li Liu 2/19/2026

Resp-Agent: An Agent-Based System for Multimodal Respiratory Sound Generation and Disease Diagnosis

Resp-Agent autonomous system with Active Adversarial Curriculum Agent for respiratory sound generation and disease diagnosis.

Ax Xianwei Mao, Kai Ye, Sheng Zhou, Nan Zhang, Haikuan Huang, Bin Li, Jiajun Bu 2/19/2026

MaS-VQA: A Mask-and-Select Framework for Knowledge-Based Visual Question Answering

MaS-VQA framework with mask-and-select approach for knowledge-based visual question answering with noisy external knowledge.

Ax Zelin Xu, Yupu Zhang, Saugat Adhikari, Saiful Islam, Tingsong Xiao, Zibo Liu, Shigang Chen, Da Yan, Zhe Jiang 2/19/2026

EarthSpatialBench: Benchmarking Spatial Reasoning Capabilities of Multimodal LLMs on Earth Imagery

EarthSpatialBench benchmarks spatial reasoning capabilities of multimodal LLMs on georeferenced Earth imagery for embodied AI systems.