TraceBench: Controlled Evaluation of LLM Agents for Time-Series Root-Cause Attribution
LLM agents are increasingly applied to anomaly detection and root-cause analysis in time-series observations collected from real-world systems; however, their performance on these tasks has not been systematically evaluated under controlled conditions. We introduce TraceBench, a simulation-based framework for generating controlled root-cause attribution tasks. In each generated task, an agent receives time-series observations produced by simulating a physical dynamical system and must determine whether a system parameter was altered during the simulation and, if so, which one. Using TraceBench
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 50%Automating root cause analysis at scale: Multi-signal correlation for cloud native incident response →
- FuzzySimilar title/name (fuzzy) · 87%NirDiamant/GenAI_Agents →
“Fuzzy title match (0.94): “TraceBench: Controlled Evaluation of LLM Agents for Time-Ser” ≈ “NirDiamant/GenAI_Agents””
- FuzzySimilar title/name (fuzzy) · 84%Unity-Technologies/ml-agents →
“Fuzzy title match (0.92): “TraceBench: Controlled Evaluation of LLM Agents for Time-Ser” ≈ “Unity-Technologies/ml-agents””
- FuzzySimilar title/name (fuzzy) · 59%datawhalechina/hello-agents →
“Fuzzy title match (0.73): “TraceBench: Controlled Evaluation of LLM Agents for Time-Ser” ≈ “datawhalechina/hello-agents””
- FuzzySimilar title/name (fuzzy) · 59%Eigenwise/atomic-agents →
“Fuzzy title match (0.73): “TraceBench: Controlled Evaluation of LLM Agents for Time-Ser” ≈ “Eigenwise/atomic-agents””
- FuzzySimilar title/name (fuzzy) · 59%jnMetaCode/agency-agents-zh →
“Fuzzy title match (0.73): “TraceBench: Controlled Evaluation of LLM Agents for Time-Ser” ≈ “jnMetaCode/agency-agents-zh””
- LinkedLinked via arxiv author · 85%Tommaso Bendinelli →
“TraceBench: Controlled Evaluation of LLM Agents for Time-Series Root-Cause Attribution”
- LinkedLinked via arxiv author · 85%Artur Dox →
“TraceBench: Controlled Evaluation of LLM Agents for Time-Series Root-Cause Attribution”