Daily updates, analyzing cutting-edge AI research papers
The paper “LLMs Improving LLMs” introduces AutoTTS, an environment where a Claude‑based explorer agent automatically searches test‑time scaling policies, achieving up to 69.5% token savings and superior accuracy on unseen models, all for $39.9 and 160 minutes without any LLM calls during evaluation.
MemEye is a multimodal memory benchmark that tests agents across eight real‑world scenarios, measuring visual evidence granularity and reasoning depth, and reveals that captions fall short for fine‑grained visual recall, highlighting the need for true visual memory in long‑term AI agents.
MemWeaver introduces a hierarchical memory that fuses behavior‑level and cognition‑level user signals, enabling large language models to generate more personalized content across multiple tasks, with extensive experiments, ablations, and an efficient incremental update mechanism demonstrating superior performance over strong baselines.
The paper introduces Safe‑SAIL, a Sparse Autoencoder Interpretation Framework for LLMs that provides pre‑explanation metrics, a segment‑level simulation to cut evaluation cost, and a 1,758‑feature safety database, enabling transparent analysis and interactive debugging of large language model safety decisions.
This survey analyzes the emerging field of Agent Skills, defining a formal skill model, categorizing acquisition pathways, detailing retrieval strategies, and outlining a five‑stage evolution process, while highlighting large‑scale skill repositories and their implications for AI product design.
The article analyzes the long‑standing privacy dilemma of cloud‑based agents, presents MemPrivacy’s three‑stage de‑identification framework and four‑level privacy taxonomy, details its two‑phase training with the MemPrivacy‑Bench dataset, and shows benchmark results where a 0.6B model outperforms GPT‑5.2 while keeping latency under 0.5 seconds.
Anthropic's two May papers reveal that simple SFT/RLHF is insufficient for safe LLMs; inserting a model‑spec mid‑training stage and synthetic‑document fine‑tuning dramatically reduces agentic misalignment, improves data efficiency, and enables models to reason about values before acting.
The paper introduces OFA-MAS, a one‑for‑all multi‑agent system that learns a universal topology designer using task‑aware graph encoding and a Mixture‑of‑Experts generator, achieving superior performance, OOD generalization, robustness, and efficiency across six major benchmarks.
SkillOS addresses the one‑off nature of current LLM agents by introducing a closed‑loop system where a trainable Skill Curator continuously extracts, updates, and manages reusable skills from execution traces, leading to measurable gains in success rates, efficiency, and cross‑task generalization.
Anthropic’s Natural Language Autoencoders (NLA) translate high‑dimensional LLM activation vectors into readable text, using an Activation Verbalizer and Reconstruction module trained via RL to maximize Fraction of Variance Explained, and reveal internal planning, language bias, tool‑call hallucinations, and hidden reasoning across multiple Claude models.
