Houyi Li

attention mechanism (1) code generation (1) low-rank approximation (1) model training (1) synthetic datum (1) scaling law (1) supervised fine-tuning (1) kv cache (1) model weight (1) data cleaning (1) code large language model (1) large language model (1) training pipeline (1) multi-matrix factorization attention (1) data-to-parameter ratio (1)