mLateOn Hits SOTA on MTEB Korean Retrieval Without Any Korean Training Data
IgorCarron · x · 2026-08-07
antoinechaffin shared test results of the mLateOn model on the MTEB Korean retrieval task. Despite not being trained on any Korean data, it achieves state-of-the-art performance, outperforming much larger models and getting close to a specialized 8B Korean model. The author notes that this impressive out-of-domain performance is exactly what late interaction mechanisms provide.
More from Models
- Report: GLM 5.5 and DeepSeek V4 Pro Launching This Month, GPT 6 Delayed — bindureddy · 2026-08-07
- ChatGPT Free Gets GPT-5.6 Luna, Paid Tiers Get Thinking Slider — testingcatalog · 2026-08-07
- Meta's Muse-Spark 1.2 Climbs to 2nd Place on SideQuest Bench — alexandr_wang · 2026-08-07
- DeepSeek v4 Flash API Pricing Revealed: Nearly 30% Cheaper — youjiaxuan · 2026-08-07
- Grok Voice's Fully In-House Stack Delivers Sub-Second Latency and Real-Time Tool Calling — Daniel_Farinax · 2026-08-07
- Professor Astonished by Gemini's Collapse as Frontier Model, Raising Enterprise Concerns — emollick · 2026-08-07