HORIZON benchmark diagnoses horizon-dependent failures in LLM agents across domains and validates a reproducible LLM-as-Judge pipeline for attributing breakdowns.
We adopt WebArena Zhou et al
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
The Long-Horizon Task Mirage? Diagnosing Where and Why Agentic Systems Break
HORIZON benchmark diagnoses horizon-dependent failures in LLM agents across domains and validates a reproducible LLM-as-Judge pipeline for attributing breakdowns.