DiagDBench: Evaluating Agentic LLMs on Interactive Database Diagnosis and Anomaly Remediation
cs.DB
Submitted: 2026-11-05
Updated: 2026-11-05
Code: https://github.com/RUCKBReasoning/DiagDBench
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- DBA-Bench: A Production-Fidelity Benchmark for LLM-Based Database Operations Agents
- LogEval: A Comprehensive Benchmark Suite for Large Language Models In Log Analysis
- DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
- DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
- ADBench: Anomaly Detection Benchmark
- Large Language Models Can Self-Improve
- SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
- Kimi K3: Open Frontier Intelligence
- Spider 2.0: Evaluating Language Models on Real-World Enterprise Text-to-SQL Workflows
- OmniSQL: Synthesizing High-quality Text-to-SQL Data at Scale
- G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment
- SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulation
- TAB: Unified Benchmarking of Time Series Anomaly Detection Methods
- SQuAD: 100,000+ Questions for Machine Comprehension of Text
- Self-Consistency Improves Chain of Thought Reasoning in Language Models
- OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments
- ReAct: Synergizing Reasoning and Acting in Language Models
- LLMIA: An Out-of-the-Box Index Advisor via In-Context Learning with LLMs
- Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena
- GaussMaster: An LLM-based Database Copilot System
Related papers
- Beyond Overlap: Hierarchical Graph Abstraction for Out-of-Core Vector Search
- EMA: Approximate Nearest Neighbor Search with General Attribute Filtering and Dynamic Updates
- Open Biomedical Knowledge Graphs at Scale: Construction, Federation, and AI Agent Access with Samyama Graph Database
- Samyama: A Unified Graph-Vector Database with In-Database Optimization, Agentic Enrichment, and Hardware Acceleration
- GenTUS: Table Union Search via Constrained Generation of Semantic Table Identifiers
- Common-Neighbor-Count-Based Representative Possible World Finding on Uncertain Graphs