VerbaNexAI at LongEval 2026 Task 4: Fully Local Temporal and Pairwise Evidence Selection for Scientific RAG

Isaac David Sánchez Sánchez Jairo Enrique S. Castañeda Edwin Puertas J. Martínez-Santos

· 1 citation · 9 references

TL;DR

The VerbaNexAI submissions to LongEval-RAG suggest that, under tight local-compute constraints, researchers should optimize evidence selection and answer style jointly rather than independently.

Semantic Scholar

View source

VerbaNexAI at LongEval 2026 Task 4: Fully Local Temporal and Pairwise Evidence Selection for Scientific RAG

Similar papers

AI_TLfanClub at FinMMEval 2026 Task 2: FiSCO-RAG for Cross-Lingual Evidence Recall in Financial Question Answering

TextSentinels @ FinMMEval 2026 Task 2: Chronological Stabilization of a BM25-Grounded Multilingual Financial Question Answering Pipeline

MG_CheckThat2026 at CheckThat! 2026: A Multilingual Cascade Pipeline for Scientific Claim Source Retrieval

The Lab Rats @ FinMMEval 2026 Task 2: LLM Translation and BAS-Tuned Chain-of-Thought for Multilingual Financial QA

UCPH-RMT at CheckThat! 2026: Evaluating Text Summaries, Atomic Facts, and Semantic Triplets as RAG Context for Fact-Checking Article Generation

SINAI-UGPLN at CheckThat! 2026: Dense Retrieval and Neural Reranking for Scientific Source Retrieval from Social Media Claims