The Morning
From the arXiv
Gubernaut: A Deterministic Homeostatic Controller for Affect-Regulated LLM Agents, Validated Across Independent Model Families
his paper introduces Gubernaut, a deterministic runtime controller for LLM agents designed to prevent reactive failures like escalation or sycophancy. It operates as a model-agnostic layer that monitors numerical telemetry (intensity, valence, repetition) and adjusts the LLM's behavior without processing text, thus creating an inherent security against injection attacks. The core contribution is a novel, robust control mechanism that ensures LLM agents maintain stable, non-reactive behavior under sustained pressure.
The Physics of Multi-Turn Long-Horizon Planning: From Pre-training to Post-training via Single- and Multi-Teacher On-Policy Agentic Distillation
This paper introduces a controlled environment to systematically study multi-turn long-horizon planning in foundation model agents. The core method involves analyzing planning ability acquisition during pre-training by manipulating data formats and introducing…
UNIFUSION: Adapting Autoregressive Language Models into Discrete Diffusion under a Unified Reverse-Rate Objective
This paper proposes UNIFUSION, a method to adapt autoregressive language models for discrete diffusion. It unifies existing diffusion objectives under a single generalized KL objective, allowing seamless switching between different corruption kernels like mask…
Kimi K3: Open Frontier Intelligence
Kimi K3 is a 2.8T parameter Mixture-of-Experts model that achieves significant scaling efficiency improvements through novel attention mechanisms (Kimi Delta Attention) and expert routing (Stable LatentMoE). Its core contribution lies in its massive scale, nat…
When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-offs
This paper systematically analyzes the trade-offs of LLM jailbreak defenses across safety, performance, and cost. It categorizes defenses by operational strategy and finds that they rarely improve downstream capabilities, instead varying in how they impact usa…
Agentic Permissions Policy Algebra for Taint Confinement in LLM Agents
This paper introduces APPA, a novel Information Flow Control framework for LLM agents. APPA addresses the usability bottleneck of traditional taint tracking by enabling engine-mana…
Beyond Aggregate Risk: Role-Stratified Conformal Risk Control for LLM Tool Calls
This paper introduces **role-stratified conformal risk control** for LLM tool calls, a method that addresses the limitations of aggregate risk control by setting separate risk budg…
ClinFusion: A Vision-Centric Multimodal LLM System for Holistic Medical Understanding
ClinFusion is a vision-centric multimodal LLM system designed for holistic medical understanding. Its core method involves a compositional and cascaded vision encoder that unifies …
D-Score: A Spectral Hidden-State Signal for Hallucination Detection in Large Language Models
This paper introduces the D-Score, a novel method for detecting hallucinations in Large Language Models. The D-Score is a spectral statistic derived from the geometry of hidden act…
DataOrchestra: Learning to Orchestrate Per-Example Curation of Pretraining Data
DataOrchestra learns to create personalized data processing pipelines for each pretraining example. It intelligently decides whether to drop, keep, or clean data, and for cleaning,…
The Town Square
AI companies are reportedly destroying rare books by using them as training data, raising concerns about the preservation of cultural heritage.
Workshops
Amnezia VPN Client is a cross-platform application enabling users to securely connect to their Amnezia VPN server on desktop and mobile devices.
This repository provides a self-hosted, AI-powered companion that can engage in real-time voice chat and play games like Minecraft and Factorio, aiming to achieve the sophistication of AI like Neuro-sama.
