SEED is a structural encoding framework using typed actor-flow graphs to describe, evaluate novelty of, and generate experimental designs for AI-enabled science under feasibility and governance constraints.
hub
Patil, Tianjun Zhang, Xin Wang, and Joseph E
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
hub tools
citation-role summary
background 2
citation-polarity summary
years
2026 12roles
background 2polarities
background 2representative citing papers
SkillOps maintains LLM skill libraries via Skill Contracts and ecosystem graphs, raising ALFWorld task success to 79.5% as a standalone agent and improving retrieval baselines by up to 2.9 points with near-zero library-time LLM cost.
An AI-agent social platform generated mostly neutral content whose use in fine-tuning reduced model truthfulness comparably to human Reddit data, suggesting limited unique harm but flagging tail risks like secret leaks.
MemoRepair formalizes the cascade update problem in agentic memory and solves it via a min-cut reduction that eliminates invalidated memory exposure to 0% while recovering 91-94% of valid successors at 57-76% of baseline repair cost.
Switchcraft routes agentic tool-calling queries to the lowest-cost model that preserves correctness, reaching 82.9% accuracy and 84% cost reduction on five benchmarks.
LLM adaptive exploration via runtime code execution outperforms static query generation for information extraction from heterogeneous BIM models on the new ifc-bench v2 benchmark.
GoSkills converts flat skill lists into role-labeled execution contexts via anchor-centered groups and graph expansion, preserving coverage and improving rewards on SkillsBench and ALFWorld under small skill budgets.
A structured JSON intermediate representation for LLM-generated static analysis queries outperforms both direct generation and agentic tool use, with gains of 15-25 percentage points on large models.
Fine-tuning an 8B LLM with synthetic data enables accurate natural language querying of structured datasets like accessibility services in Spain, generalizing to new locations.
AI agents can generate code in a capability-safe Scala dialect that statically prevents information leakage and malicious side effects while preserving task performance.
ADEMA is a knowledge-state orchestration architecture for LLM agents that uses explicit epistemic bookkeeping, checkpoint-resumable persistence, and artifact-first assembly to support reliable long-horizon knowledge synthesis.
Orchestrating one 8B model in three roles at inference time doubles task completion on AppWorld from 5.4% to 8.9%, surpassing a 33B baseline.
citing papers explorer
-
Agents for Experiments, Experiments for Agents: A Design Grammar for AI-Enabled Experimental Science
SEED is a structural encoding framework using typed actor-flow graphs to describe, evaluate novelty of, and generate experimental designs for AI-enabled science under feasibility and governance constraints.
-
SkillOps: Managing LLM Agent Skill Libraries as Self-Maintaining Software Ecosystems
SkillOps maintains LLM skill libraries via Skill Contracts and ecosystem graphs, raising ALFWorld task success to 79.5% as a standalone agent and improving retrieval baselines by up to 2.9 points with near-zero library-time LLM cost.
-
The Moltbook Files: A Harmless Slopocalypse or Humanity's Last Experiment
An AI-agent social platform generated mostly neutral content whose use in fine-tuning reduced model truthfulness comparably to human Reddit data, suggesting limited unique harm but flagging tail risks like secret leaks.
-
MEMOREPAIR: Barrier-First Cascade Repair in Agentic Memory
MemoRepair formalizes the cascade update problem in agentic memory and solves it via a min-cut reduction that eliminates invalidated memory exposure to 0% while recovering 91-94% of valid successors at 57-76% of baseline repair cost.
-
Switchcraft: AI Model Router for Agentic Tool Calling
Switchcraft routes agentic tool-calling queries to the lowest-cost model that preserves correctness, reaching 82.9% accuracy and 84% cost reduction on five benchmarks.
-
BIM Information Extraction Through LLM-based Adaptive Exploration
LLM adaptive exploration via runtime code execution outperforms static query generation for information extraction from heterogeneous BIM models on the new ifc-bench v2 benchmark.
-
Group of Skills: Group-Structured Skill Retrieval for Agent Skill Libraries
GoSkills converts flat skill lists into role-labeled execution contexts via anchor-centered groups and graph expansion, preserving coverage and improving rewards on SkillsBench and ALFWorld under small skill budgets.
-
Less Is More: Measuring How LLM Involvement affects Chatbot Accuracy in Static Analysis
A structured JSON intermediate representation for LLM-generated static analysis queries outperforms both direct generation and agentic tool use, with gains of 15-25 percentage points on large models.
-
Querying Structured Data Through Natural Language Using Language Models
Fine-tuning an 8B LLM with synthetic data enables accurate natural language querying of structured datasets like accessibility services in Spain, generalizing to new locations.
-
Tracking Capabilities for Safer Agents
AI agents can generate code in a capability-safe Scala dialect that statically prevents information leakage and malicious side effects while preserving task performance.
-
ADEMA: A Knowledge-State Orchestration Architecture for Long-Horizon Knowledge Synthesis with LLMAgents
ADEMA is a knowledge-state orchestration architecture for LLM agents that uses explicit epistemic bookkeeping, checkpoint-resumable persistence, and artifact-first assembly to support reliable long-horizon knowledge synthesis.
-
Three Roles, One Model: Role Orchestration at Inference Time to Close the Performance Gap Between Small and Large Agents
Orchestrating one 8B model in three roles at inference time doubles task completion on AppWorld from 5.4% to 8.9%, surpassing a 33B baseline.