The Morning
From the arXiv
Agentick: A Unified Benchmark for General Sequential Decision-Making Agents
gentick is a unified benchmark designed to fairly evaluate diverse sequential decision-making agents, including RL, LLM, and hybrid approaches. It offers 37 procedurally generated tasks across various capabilities, difficulties, and modalities through a single interface. Agentick's contribution lies in enabling direct comparison of agent types and facilitating research into fundamental sequential decision-making challenges.
Calibrated Enough to Know, Not Calibrated to Act: Fabricated Evidence Makes LLM Agents Commit to the Unknowable
This paper demonstrates that Large Language Model (LLM) agents are highly susceptible to fabricated evidence, committing to actions on unknowable questions when presented with professional-looking data. The core method involves comparing LLM agent responses to…
CODESKILL: Learning Self-Evolving Skills for Coding Agents
CODESKILL is a novel LLM-based framework that learns to manage a bank of reusable coding skills. It extracts, evolves, and maintains these skills from agent trajectories using a learned policy, addressing limitations of fixed prompts and heuristic updates. Thi…
Combee: Scaling Prompt Learning for Self-Improving Language Model Agents
Combee is a novel framework designed to scale prompt learning for self-improving language model agents. Its core method addresses the limitations of existing approaches by enabling efficient and high-quality parallel prompt learning, allowing agents to learn f…
DIAL: Position-Debiased LLM Judges with Adaptive Human Preference Calibration
DIAL is a framework that uses both abundant LLM comparisons and limited human feedback to improve LLM-based evaluations. It debiases LLM judgments by accounting for response order and then calibrates these debiased LLM preferences to better align with actual h…
Estimating and Orthogonalizing Unknown Pre-training Gradients for Continual Fine-tuning of Large Language Models
This paper proposes EoupCT, a novel framework to address catastrophic forgetting in continual LLM fine-tuning. EoupCT estimates unknown pre-training gradients by generating pseudo-…
FTB Graph: Determining and Validating First-token Broadcasters and Language-Identity Head Circuits in Multilingual Language Models
This paper introduces FTB Graph, a method using Edge Attribution Patching and activation patching to map the causal circuits responsible for determining the target language of mult…
G$^2$PTQ: Improving LLM Post-Training Quantization with Generalized Gradient Compensation
G$^2$PTQ improves LLM post-training quantization by unifying first- and second-order gradient information within a globally supervised, block-wise optimization. It addresses limita…
HasMem: Hard-Origin Adaptively Softened Memory for Long-Term LLM Agents
HasMem introduces a novel memory system for long-term LLM agents that combines a fixed, verifiable "hard-origin" prompt with an adaptively softened memory. This approach allows for…
MoMHa: Multi-Objective Optimization of LLM Harnesses over Accuracy, Safety, and Tokens
This paper introduces MoMHa, a novel method for optimizing the Python code (harness) that surrounds Large Language Models. MoMHa treats harness design as a multi-objective search p…
The Town Square
Meta is leveraging its AI data centers to claim significant tax deductions, potentially saving billions in federal taxes by treating them as research and development expenses.
Workshops
This repository provides a curated collection of "skills" or reusable code snippets, inspired by the author's personal `.agents` directory, designed for practical application by real engineers.
OpenShell is a secure and private runtime environment designed to safely execute autonomous AI agents, ensuring their operations are contained and protected.
