Improving Instruction Tuning for LLMs: Meta AI Presents the OPT-IML Benchmark of 2000 NLP Tasks | Synced
In the new paper OPT-IML: Scaling Language Model Instruction Meta Learning through the Lens of Generalization, a Meta AI research team presents OPT-IML Bench, an Instruction Meta Learning benchmark...
Source: Synced | AI Technology & Industry Review
In the new paper OPT-IML: Scaling Language Model Instruction Meta Learning through the Lens of Generalization, a Meta AI research team presents OPT-IML Bench, an Instruction Meta Learning benchmark comprising 2000 NLP tasks and an evaluation framework for model generalization.