💾 CPU Caching - emschwartz · Scour

Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs 🕳LLM Vulnerabilities

Machine learning-driven alignment architecture of heterogeneous data with transient varying semantics 🏗️LLM Infrastructure

easyaligner: Forced Alignment Made Easy 🔤Tokenization

kb-labb.github.io·6d

Takes on Automating Alignment 👨‍💻AI Coding

lesswrong.com·4d

polish by pbakaus/impeccable 🎨Design Tokens

CAP: Controllable Alignment Prompting for Unlearning in LLMs 🪄Prompt Engineering

Alignment by Default? 🛡️AI Safety

blog.cosmos-institute.org·4d·Hacker News

ReCAPA: Hierarchical Predictive Correction to Mitigate Cascading Failures 🔄Incremental Computation

Monday AI Radar #22 🆕New AI

lesswrong.com·3d

Value-Conflict Diagnostics Reveal Widespread Alignment Faking in Language Models 🛡️AI Safety

From Noise to Intent: Anchoring Generative VLA Policies with Residual Bridges 🧠LLM Inference

Alignment has a Fantasia Problem 🪄Prompt Engineering

MGDA-Decoupled: Geometry-Aware Multi-Objective Optimisation for DPO-based LLM Alignment ⚡PGO

Sensitivity Uncertainty Alignment in Large Language Models 🔤Tokenization

VG-CoT: Towards Trustworthy Visual Reasoning via Grounded Chain-of-Thought ✨Gemini

C-Mining: Unsupervised Discovery of Seeds for Cultural Data Synthesis via Geometric Misalignment 🐘pgvector

ONOTE: Benchmarking Omnimodal Notation Processing for Expert-level Music Intelligence ✨Gemini

Harmful Intent as a Geometrically Recoverable Feature of LLM Residual Streams 🛡️AI Safety

Continual Safety Alignment via Gradient-Based Sample Selection 🛡️AI Safety

When Choices Become Risks: Safety Failures of Large Language Models under Multiple-Choice Constraints 🛡️AI Safety

Log in to enable infinite scrolling