Revision History
Multi-token prediction · 4 revisions
Sizes are character counts of the article source. The signed number is the change from the previous revision.
Recent edit summaries
Detailed summaries recorded by editors. Generic maintenance summaries are omitted here; every recorded revision remains below. Dates describe the edit, not necessarily the event it covers.
- Correction: fixed misattributed figures and references (llama.cpp, Qwen3-Next, FastMTP authors, MTP-D source, blockwise decoding speedup, LM-head size, drafter storage and verification cost claims); added section on deploying MTP layers as speculative drafters with MoE verification costs and the NVIDIA DGX Spark Ling-3.0-flash example
Version 4 · Oct 29, 2026, 09:33 PM
- Improve internal links (2026-10-13)
Version 3 · Oct 13, 2026, 10:03 AM
- Added 6 contextual internal links
Version 2 · Jul 23, 2026, 03:31 PM