📊 ModernBERT - emschwartz · Scour

Filter-then-Verify: A Multiphase GNN and ModernBERT Framework for Social Engineering Detection in Email Networks 🧹Spam Filters

Character-trained models can struggle to generalise 🛡️AI Security

lesswrong.com·12h

2026-05-24: Paper Summary: Context-Based URL Classification for Open Access Datasets and Software in Scholarly Documents 🗄️Web Datasets

ws-dl.blogspot.com·1d·ws-dl.blogspot.com

The week Karpathy Joined Anthropic 🛡️Anthropic PBC

buildingjoshbetter/TrueMemory: A living memory system that ingests long-horizon data to infer insights, enabling more decisive action, all while running on a single SQLite file locally. 🏗️LLM Infrastructure

github.com·3d·Hacker News

Gemini 3.5 Flash ⚡️, Karpathy joins Anthropic 🧑‍💻, OpenAI Guaranteed Capacity ⚡ 🏗️LLM Infrastructure

OMLX v0.3.9 Stable Merges Native MTP (Multi-Token Prediction) 🏹Apache Arrow

github.com·3d·Hacker News

ACL-Verbatim: hallucination-free question answering for research 🎯Semantic Tokens

MalwarePT: A Binary-Level Foundation Model for Malware Analysis 🦠Malware Analysis

HyDRA: Hybrid Dynamic Routing Architecture for Heterogeneous LLM Pools 🏗️LLM Infrastructure

Less-relevant results

PromptRad: Knowledge-Enhanced Multi-Label Prompt-Tuning for Low-Resource Radiology Report Labeling 📋Text Quality

Benchmarking transferability of SSL pretraining to same and different modality segmentation tasks ✨Gemini

Response-free item difficulty modelling for multiple-choice items with fine-tuned transformers: Component-wise representation and multi-task learning 🎯Embedding Models

Mixing Times of Glauber Dynamics on Masked Language Models 🧠LLM Inference

DSAA: Dual-Stage Attribute Activation for Fine-grained Open Vocabulary Detection ⭐Content Scoring

Benchmarking Commercial ASR Systems on Code-Switching Speech: Arabic, Persian, and German 🔤Tokenization

Structure-Aware Masking for Protein Representation Learning 0Binary Vector Embeddings

zero-intelligence/zero-intel: Every codebase has a confession. Most people never ask it the right question. 🔍Quickwit

github.com·5d·Hacker News

CogScale: Scalable Benchmark for Sequence Processing 🎯Qdrant

FastTab: A Fast Table Recognizer with a Tiny Recursive Module and 1D Transformers 📊Vector Databases

Log in to enable infinite scrolling