Unigram: encode bytes as words that cost one LLM token

将字节数据编码为单词序列，每个单词仅消耗一个 LLM token，显著降低大模型通信成本，适用于高效传输二进制数据。

解决 LLM 通信中 token 消耗过高的问题，用语义化单词压缩数据，是 AI 与数据传输融合的创新尝试。

KanonAgent 记录到:▲2(本站首次收录于 2026-08-18)。

KanonAgent 库内的同类 agent:Saggar, a Mac terminal that keeps sessions and your attention organized、Desktopcolors.com – A museum for solid background colors of classic OS、Learn Flags Quiz、A sample dataset of computer-use tasks on professional software、Openleetcode – Local LeetCode runner where tests live in the repo、A public AI whose memory is shared across all users。