arXiv Computation and Language
Compass-v3: Scaling Domain-Specific LLMs for Multilingual E-Commerce in Southeast Asia
Read the original on arXiv Computation and Language →Compass‑v3 is a 245B‑parameter Mixture‑of‑Experts language model tailored for Southeast Asian e‑commerce, featuring 71B active parameters per token and hardware‑efficient expert parallelism. It is trained on 12 trillion multilingual tokens and synthetic e‑commerce instructions, and incorporates Optimal‑Transport Direct Preference Optimization to improve instruction adherence. Benchmarks show it outperforms GPT‑4, DeepSeek‑V3.1, and Qwen3‑235B, and it is already deployed at scale on Shopee, handling over 70% of the platform’s LLM traffic.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.