CascadeMind at SemEval-2026 Task 4: A Hybrid Neuro-Symbolic Cascade for Narrative Similarity

TL;DR

CascadeMind combines confidence-based routing of multiple LLM votes with symbolic reasoning to improve narrative similarity detection, emphasizing the importance of targeted computation over auxiliary representations.

Contribution

Introduces a hybrid neuro-symbolic cascade system that dynamically allocates resources based on vote confidence, achieving competitive results in narrative similarity tasks.

Findings

Supermajority votes achieve 85% accuracy

Split votes reach 67% accuracy

Symbolic ensemble adds negligible end-to-end gain

Abstract

Across self-consistency samples from an LLM, vote agreement tracks instance difficulty: on SemEval-2026 Task 4 (Narrative Story Similarity), supermajority cases (>= 7/8 votes) resolve at 85 percent accuracy, split votes at 67 percent, and perfect ties at 61 percent, a monotone gradient that holds across the development set. We exploit this in CascadeMind, which routes eight Gemini 2.5 Flash votes by consensus, escalates split votes to additional sampling rounds, and falls through to a symbolic ensemble of theory-inspired narrative signals only on perfect ties (5 percent of cases). The system reached 72.75 percent on Track A test, placing 10th of 44 teams. Ablations show that the symbolic component contributes negligibly end-to-end and that nearly all gains come from confidence-aware routing. The takeaway is methodological: for narrative similarity, calibrating when to spend more compute on a hard instance matters more than adding auxiliary representations to reason about it.