FilBBQ provides a culturally adapted Filipino bias benchmark for QA models plus a multi-seed evaluation protocol that detects sexist and homophobic biases while showing score variability across runs.
Title resolution pending
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CL 3representative citing papers
Ontology-based constraints combined with hybrid fine-tuning enable consistent control over LLM conversational outputs on proficiency and polarity tasks, outperforming baselines across seven models.
SETUP achieves AnCast 84 and SMATCH++ 91 on English-to-UMR parsing by fine-tuning AMR models and converting from Universal Dependencies, outperforming prior baselines.
citing papers explorer
-
Robust Bias Evaluation with FilBBQ: A Filipino Bias Benchmark for Question-Answering Language Models
FilBBQ provides a culturally adapted Filipino bias benchmark for QA models plus a multi-seed evaluation protocol that detects sexist and homophobic biases while showing score variability across runs.
-
Conversational Control with Ontologies for Large Language Models: A Lightweight Framework for Constrained Generation
Ontology-based constraints combined with hybrid fine-tuning enable consistent control over LLM conversational outputs on proficiency and polarity tasks, outperforming baselines across seven models.
-
SETUP: Sentence-level English-To-Uniform Meaning Representation Parser
SETUP achieves AnCast 84 and SMATCH++ 91 on English-to-UMR parsing by fine-tuning AMR models and converting from Universal Dependencies, outperforming prior baselines.