fix: initialize embeddings_pre_norm_masked=false in llama_context#23256

ggerganov merged 1 commit into

ggml-org:masterggml-org/llama.cpp:masterfrom

abetlen:fix/qwen35-pre-norm-mask-initabetlen/llama.cpp:fix/qwen35-pre-norm-mask-initCopy head branch name to clipboard