The Morning
From the arXiv
Null-Space Constrained Low-Rank Adaptation for Response-Specified Large Language Model Unlearning
his paper introduces Null-Space Constrained Response-Specified Unlearning (NSRU), a method for unlearning specific knowledge from large language models while preserving general capabilities. NSRU achieves this by projecting low-rank adaptation updates into the null space of learned "retain" subspaces, ensuring that modifications are localized and do not disrupt benign knowledge. The core contribution is a projection-constrained framework that enables precise control over unlearning by specifying safe target responses and confining updates to preserve existing functionalities.
ReasonAlloc: Hierarchical Decoding-Time KV Cache Budget Allocation for Reasoning Models
ReasonAlloc addresses the KV cache bottleneck in LLM reasoning by dynamically allocating its budget. It uses a hierarchical approach: an offline strategy identifies an architecture-driven "Reasoning Wave" of demand across layers, and an online strategy realloc…
Role-Agent: Bootstrapping LLM Agents via Dual-Role Evolution
Role-Agent addresses LLM agent limitations by using a single LLM as both agent and environment for bootstrapped co-evolution. It employs World-In-Agent (WIA) for environment-aware reasoning via state prediction rewards, and Agent-In-World (AIW) to refine train…
Workflow-GYM: Towards Long-Horizon Evaluation of Computer-use Agentic tasks in Real-World Professional Fields
Workflow-GYM is a new benchmark designed to evaluate AI agents' ability to perform complex, long-horizon tasks within professional software environments, unlike existing benchmarks that focus on simpler applications. Its core contribution is to highlight the s…
Flow-DPPO: Divergence Proximal Policy Optimization for Flow Matching Models
Flow-DPPO addresses limitations of existing RL methods for flow matching models by replacing noisy policy ratio clipping with a divergence proximal constraint. Its core method leverages the Gaussian nature of per-step policies in flow models to efficiently com…
Does Reasoning Preserve Alignment? On the Trustworthiness of Large Reasoning Models
This paper investigates whether training large language models for reasoning sacrifices their alignment with human values like safety and bias avoidance. The study finds that curre…
It Takes One to Bias Them All: Breaking Bad with One-Shot GRPO
This paper demonstrates that a single biased example, using Group Relative Policy Optimization (GRPO), is enough to systematically bias large language models. This bias then genera…
Pushing the Limits of LLM Tool Calling via Experiential Knowledge Integration and Activation
This paper addresses LLM tool-use limitations by integrating and activating experiential knowledge. The core method involves acquiring instance-level knowledge, which proves more e…
A History-Aware Visually Grounded Critic for Computer Use Agents
This paper introduces HiViG, a novel framework for improving computer use agents. HiViG addresses limitations of existing critics by incorporating both a history of past actions an…
AuRA: Internalizing Audio Understanding into LLMs as LoRA
AuRA internalizes audio understanding into Large Language Models (LLMs) by distilling audio encoding capabilities into a lightweight LoRA adaptation. It trains the LLM to mimic the…
The Town Square
The article argues that CEOs viewing AI as a direct replacement for employees are demonstrating poor leadership, as AI should be used to augment human capabilities, not eliminate them.
Workshops
This repository offers a marketplace of over 100 agentic skills, commands, and plugins designed to support the entire product management lifecycle, from initial discovery and strategy to execution, launch, and growth.
Tolaria is a desktop application designed to efficiently manage markdown knowledge bases, offering a user-friendly interface for organizing, creating, and accessing your notes.
