Newest papers

The Instruction Tax on Factual Recall: Base Models Can Beat Their Instruction-Tuned Counterparts

Christophe Khalil, Aya Mourad · 2026 · arXiv

DOI: 10.48550/arXiv.2610.28407 · open-access PDF

Portuguese

Apresentamos a nossa solução à AKBC LM-KBC 2026 Shared Task, onde alcançamos uma pontuação de 0,6836 no tabuleiro oficial, mais do que o dobro da linha de base de 0,2964. Esta tarefa requer a geração de todos os possíveis objetos para pares de relacionamento de assunto usando um modelo de idioma com até 32 bilhões de parâmetros, sem recuperação ou ajuste fino envolvido. Mostramos que a metodologia de prompting é tão importante quanto a arquitetura do modelo: enquanto a versão do modelo ajustado com instruções atinge 0,5681 com sua interface de bate-papo, o modelo de base atinge 0,6836 quando solicitado usando a conclusão de texto.

Machine translation (Meta M2M100); the English original follows.

Abstract

We present our solution to the AKBC LM-KBC 2026 Shared Task, where we achieved a score of 0.6836 on the official leaderboard, more than double the baseline of 0.2964. This task requires generating all possible objects for subject-relation pairs using a language model with up to 32 billion parameters, with no retrieval or fine-tuning involved. We show that the prompting methodology is as important as the model architecture: while the instruction-tuned model version achieves 0.5681 with its chat interface, the base model achieves 0.6836 when prompted using text completion. Thus, instruction tuni

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.28407, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange