t-mem
t-mem is T-Mem: Memory That Anticipates, Not Archives — accepted to EMNLP 2026 Main Conference. Trigger-augmented graph memory for long-term conversational QA. - Sherlockwz/T-Mem. It is ranked #62 on the Dev Radar, in Other, first seen 1 days ago and shared in 4 posts (81.5k views).
GitHub - Sherlockwz/T-Mem: T-Mem: Memory That Anticipates, Not Archives — accepted to EMNLP 2026 Main Conference. Trigger-augmented graph memory for long-term conversational QA. · GitHub Skip to content Navigation Menu Sign in Appearance settings Search / Sign in Sign up Appearance settings You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert {{ message }} Sherlockwz / T-Mem Public Notifications You must be signed in to change notification settings Fork 3…
What people said about t-mem on X
This might be the first open-source memory architecture to cover associative recall. The rest stop at similarity search. (T-Mem, EMNLP 2026) Most memory systems retrieve by similarity. T-Mem predicts when a memory will matter, before you ask. → write: memories ship with a trigger at store time, basically "when will…
— @TencentAI_News, 1 days ago · 508 likes · see the post
腾讯 PCG 团队开源长期记忆系统 T-Mem，论文已被 EMNLP 2026 主会接收。它会在记住一件事时，顺手记录「这件事以后可能在什么场景下有用」。 比如用户曾说「同事海鲜过敏」，后来只问「团建去哪吃」。两句话几乎没有共同关键词，普通记忆系统可能找不到这条旧信息。T-Mem 会提前把「海鲜过敏」和「聚餐选餐厅」关联起来，之后遇到类似场景也能把它找回来。 评测显示，普通记忆系统一旦失去明显的关键词提示，表现会明显变差；T-Mem 受影响更小。把负责提前预测未来场景的 Horizon Trigger 去掉后，T-Mem 的表现也会明显下降，说明这套「提前埋线索」的机制确实起了作用。
— @0xLogicrw, 21 h ago · 126 likes · see the post
Most AI memory systems answer the wrong question. They're built to retrieve the memory most similar to what you just asked. The harder question is which memory will matter later, before anyone asks anything. That's the gap a new Tencent paper, T-Mem, is aimed at. Most memory systems, from vector stores to the newer…
— @stretchcloud, 15 h ago · 2 likes · see the post
Alternatives to t-mem
- Spark-X2.5 — Model Scope
- qwen-image-2.1 — Qwen's most powerful open-source image generation model - QwenLM/Qwen-Image-2.1
- mimo-v2.6-flash-rl — We’re on a journey to advance and democratize artificial intelligence through open source and open science.
- glm-5.3-flash-exl3-2x-dgx-sparks — GLM-5.3 Flash EXL3 for 2-4x DGX Sparks. Contribute to MiaAI-Lab/GLM-5.3-Flash-EXL3-2x-DGX-Sparks development by…
- mimo-v2.6-pro-rl — We’re on a journey to advance and democratize artificial intelligence through open source and open science.
- Ling-3.0-flash-Fin — Qwen-Drive-1.0-4B is a vision-language foundation model that handles 3D perception, driving VQA, and motion planning…
t-mem in numbers
- Rank on the Dev Radar: #62 of 2352
- Shared in 4 posts by 4 accounts: @TencentAI_News, @0xLogicrw, @stretchcloud, @KakaluoteW45042
- 81.5k views on those posts
- First seen 1 days ago, last shared 15 h ago
- Pricing seen by Jev: open source
- Market: Other
FAQ
What is t-mem?
T-Mem: Memory That Anticipates, Not Archives — accepted to EMNLP 2026 Main Conference. Trigger-augmented graph memory for long-term conversational QA. - Sherlockwz/T-Mem It was first shared on X 1 days ago and is ranked #62 on the Dev Radar.
Is t-mem free?
It is open source.
Who shared t-mem?
4 accounts on X, including @TencentAI_News, @0xLogicrw, @stretchcloud, in 4 posts totalling 81.5k views.
Every post is read and classified by Jev (TypeSafe): what it is, which market it belongs to, and whether the link is a real tool. 20.7k posts from 4.9k X accounts over the last 21 days, 2.4k tools, 12 markets. Collected every 5 minutes, fully re-ranked every hour — last update 2026-09-22 08:21 UTC. Full method.