paper analysis
8 notes, newest first.
ReguSim: Evaluating LLM Agent Rule Grounding in Financial Compliance
The paper introduces ReguSim, a controlled environment for evaluating Language Learning Models (LLM) in financial compliance. It presents ReguBench, a monitoring benchmark, to analyze four artifacts: reasoning…
Training-Free Inference-Time Self-Reflection and Cost-Bounded Early Stopping for Large Language Models
The paper presents EvoResearcher, a training-free protocol for adding cost-bounded self-reflection to large language models (LLMs). The protocol is designed to work with a single frozen LLM backbone and aims to improve…
CACHE-UK: A Stability-Aware Memory Editor for Sequentially Updated Quantized LLMs in Finance
Researchers introduce CACHE-UK, a framework designed for efficiency and accuracy in deploying large language models (LLMs) in the dynamic field of finance. The system addresses the 'quantization stability crisis' that…
Learning to Trace Seiberg Dualities
In this paper, the authors use machine learning methods to develop a practical tool for establishing Seiberg dualities in supersymmetric quiver gauge theories. They find that network architectures consisting of…
MemOps: Benchmarking Lifecycle Memory Operations in Long-Horizon Conversations
The paper introduces MemOps, a new benchmark for evaluating the lifecycle of memory operations in long-horizon conversations for LLM-based agents. Unlike existing benchmarks that focus on final answers, MemOps focuses…
GitLake: Git-for-data for the agentic lakehouse
The paper introduces GitLake, a system designed for managing data in an agent-first lakehouse using Git principles. It allows agents to work on isolated branches while humans review changes, and ensures all outputs are…
AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents
The researchers introduce a new contract system for long-term learning machine (LLM) agents that limits the memory of each decision to a fresh user message, assembled by typed retrieval. They test this contract in the…
SkillFuzz: Fuzzing Skill Composition for Implicit Intents Discovery in Open Skill Marketplaces
The paper proposes SkillFuzz, a method to detect potential unintended objectives (implicit intents) in Large Language Model (LLM)-based agents that use reusable skills from open skill marketplaces. SkillFuzz identifies…
