포도_podo

[Big Mayo @Big_Mayo_]

Lucky Archive

[에스티미디어(ST Media) @STMEDiA_info]

오랫동안 기다려주신 여러분, 감사합니다.

이제 New Universe라는 이름으로, 유니(UNI)와 새 캐릭터들이 함께 만들어갈 새로운 여정을 여러분과 함께 시작하고 싶습니다.

자세한 소식은 아래에서 확인해 주세요.

https://t.co/ltWgrtsNDT

#유니 #UNI #NewUniverse #음성제공자모집

[Katie Paxton-Fear @InsiderPhD]

🧵Can we trust Chinese open weight models? Was a question a lot of people asked after GLM 5.2 was released, scoring very well on coding benchmarks, and suspiciously Claude-like. So I turned an open-weight coding model into a backdoor with 1hr and <$100. Let's talk about it

[株式会社なのです @nanodes3d]

「先生、ちっと時間もらってくぜ！」にて登場したネルの3Dモデルを制作させて頂きました。 #ブルアカ #ちっと時間もらってくぜ

[🌟각설이🍷🩸 @ga_gs0ise]

#Lize_live

방송을 계속 하지 못했을수도 있있던 썰

계속 스텔라이브에 있어줘서 고마워...

[Avi Chawla @_avichawla]

NVIDIA researchers built a new transformer variant.

One small change to the layers made:

- decoding 1.7x faster
- long-reasoning accuracy up 6.5 points

In a typical transformer architecture, every attention layer computes Q, K, and V.

NVIDIA's tweak adds a fourth projection, which predicts what the next layer will need.

To understand why they did this, let's first see what happens in a Transformer architecture during inference right now.

Sparse attention was an attempt to handle long-context inference. Instead of attending to every cached token, modern designs score the KV cache in blocks, keep the top-k, and attend only to those.

That cuts attention compute and bandwidth, but this still leaves us with two problems.

> First, the KV cache still grows with every generated token.

At 100K+ context, it no longer fits in GPU memory and gets offloaded to CPU RAM.

Now every layer must first copy its selected KV blocks from CPU memory back to the GPU. That copy is slow, the GPU sits idle while it waits, and the stall repeats at every layer of every decode step.

> Second, the selection step itself is not free.

Standard selectors score every candidate block with every query head in a GQA group (grouped-query attention, where several query heads share one KV head), then softmax each head's scores and sum them across the group.

During decode, the sparse attention itself is cheap because there is only one query token.

But the expensive part is deciding which blocks to attend to, and that cost keeps growing with context length.

Both problems trace back to the same design in today's sparse attention methods, i.e., the attention query drives the block selection.

Selection needs the query vector Q, and Q only exists once its layer is already running. By then, it's too late to fetch anything early.

The query also drags its multi-head layout into selection, so all that scoring computation runs just to make one top-k decision.

A recent paper from NVIDIA and MIT called SparDA breaks this coupling with one architectural change.

Each layer now emits four projections instead of three:

↳ Q, K, V, and a Forecast.

The Forecast from layer L predicts which KV blocks layer L+1 will need.

Layer L+1's own query performs the sparse attention over those selected blocks.

This one change fixes both problems.

Since the next layer's block set is known while the current layer is still computing, the runtime fetches those blocks from CPU memory on a separate CUDA stream.

The copy overlaps with the current layer's compute, so the GPU no longer waits for it.

And since the Forecast is separate from the attention query, it doesn't need one score per query head.

SparDA uses one Forecast head per GQA group, which removes the per-query-head scoring loop and skips the softmax step entirely.

DeepSeek did something similar in DSA, where a small indexer picks important tokens instead of the query doing it.

SparDA applies the same idea to blocks and adds the prefetch angle that DSA doesn't touch.

The cost of the change is small.

The Forecast adds just 33.5M parameters on an 8B model (0.41%), and only those projections are trained, using a KL loss that matches the original selector's block distribution.

On MiniCPM4.1-8B and NOSA-8B, accuracy matches or beats the sparse baseline, with NOSA-8B gaining +6.5 on long reasoning.

Prefill runs up to 1.25x faster and decode up to 1.7x faster than the sparse offload baseline.

There's one more benefit.

Because prefetch hides the offload cost, most of the KV cache can live in CPU RAM, and the freed GPU memory fits much bigger batches, pushing decode throughput up to 5.3x over the non-offload sparse baseline.

That said, this lookahead will only pay off during decode with CPU offload. During prefill, all keys already live on the GPU, so the gain there comes purely from the cheaper selection.

Here's the paper: https://t.co/aOszIa8YDZ

I wrote a first-principles breakdown of how the KV cache works. It walks through why the model stores keys and values at all, why the cache grows with every token, and a comparison of LLM generation speed with and without KV caching.

Read it below.

[ぷ。 @PpFjvDJUdK54073]

日中韓、擬人化してみた
#国擬人化
#擬人化

[US ASSAULT @USASSAULTBF4]

Happy 4th of July folks! 🇺🇸🫡

[마법쨔 @Magicjjang1011]

토스에서 마크 서버만들수있네
도대체 뭐지
일단 해보니까 잘 되긴하네

[무슨 일이 일어나고 있나요? @museun_happen]

홈플러스 회생절차 폐지, 파산 확정

법원에서 회생도 불가능하다 판단하고 전부 폐점행 예정

[Kizuna AI / キズナアイ @aichan_nel]

みんなからのリプやコメントみてたらこんな時間に！！！！！！！

改めまして！！！
#AIParty2026 心からありがとうございました🎀✨

ゲストのみんなとの楽しいゲームからはじまって！
大切に作ってきた新曲や新衣装もお披露目できて！
短かったけどみんなで必死に準備してきたライブもできて！

キズナアイ愛されているなぁって
実感できた素敵な誕生日でした🎉🎉🎉

もう10歳！でもまだ10歳！
今年からはじまる10周年に向けてどんどん盛り上げていきたいです！！！

でもそれには本当にみんなの力が必要で
みんながいるからつながっていけるって思ってます🎀✨

これからもずーーーーっとよろしくね！！
みんな大好きლ(´ڡ`ლ)

[Kizuna AI Inc. @KizunaAIinc]

／
Kizuna AI WORLD TOUR
GO TO 2027🌍✈️
＼

今年12月1日に10周年を迎える「Kizuna AI」🎀
記念すべき10周年イヤーを盛り上げる、ワールドツアー企画始動！

世界中のみんなに会いに行きます！
より多くの場所でライブをやっていきます✨
世界中から出演オファーお待ちしております🫶

詳細はこちら⏬
https://t.co/FcY5JXuAbi

🎀出演依頼はこちら🎀
https://t.co/SK30wuRIrP

#KizunaAI #GoTo2027

[Shylily @shylilytwitch]

⚠️ FURRY MODEL ⚠️

Deep down inside I've always been a feral furry 😼 🐬

4th of July
11AM PDT
2PM EDT
8PM CEST

On Twitch 💯

[Bad Blue Archive Memes @CryingEmojiman]

Torture

[田中ヒメ(HIMEHINA)🥕 @HimeTanaka_HH]

初ソウルワンマン最高だったあああ❣️ʚ̴̶̷᷄＿ʚ̴̷̷᷅❤️‍🔥❤️‍🔥
みんなの熱気ｱﾁｱﾁ声ﾃﾞｶﾃﾞｶでソウル(魂)感じたぜぃっ‼️🔥
難しいのに日本語でいっぱい歌ってくれてありがとう😭感動‼️‼️
また絶対帰ってくるから会いに来てねーっ！！！ᐡ⸝⸝> ·̫ <⸝⸝ᐡ💗
#HIMEHINAWorldTourSeoul

[Bad Blue Archive Memes @CryingEmojiman]

She is the best

[모래늑대🐺 @moraeneukdae]

젠레스 존 제로 입장모션 패러디... 학생들의 각오를 다지거나 여유있는 미소에 매료되네...

[常闇トワ👾 @tokoyamitowa]

お世話になっております。
常闇トワです。

まずはご心配をおかけしてしまい申し訳ございません。

6月21日の朝、突然強いめまいと立ちくらみの症状が出てしまい、自力で立つことや歩くことができなくなったため救急搬送され治療を受けておりました。

病院へは白上フブキちゃんについてきてもらい、その後担当マネージャーにも連絡・対応・駆けつけてもらいました。本当に感謝しています。

その後、セカンドオピニオンも含めて診察を受けた結果、現時点ではメニエール病の疑いがあるとの診断を受けています。

医師からは今月いっぱい休養するよう指示を受けており、会社とも相談した結果、今月は活動をお休みさせていただくことになりました。
来月以降については、今後の体調や診察結果を見ながら判断していきます。

また、V最スト6については当日の辞退となってしまい、主催の渋谷ハルさん、チームメンバーの皆さん、応援してくださっていた皆さん、そして関係者の皆様に多大なご迷惑をおかけしてしまい、本当に申し訳ありませんでした。

最後まで出場したい気持ちはありましたが、その時の状態では試合に参加できる状況ではありませんでした。

現在は薬の効果もあり、自力で歩けるところまで回復しています。
まだ頭を動かした際のめまいやふらつきは残っていますが、少しずつ体を動かして良くなっています。

命に別状はありませんので、その点は安心してください！

まずはしっかり休んで治療に専念し、元気な姿で戻ってこられるよう頑張ります🙇‍♂️

たくさんの温かいメッセージやお気遣い、本当にありがとうございます。

改めて、ご心配とご迷惑をおかけして大変申し訳ございませんでした。

[Sakana AI @SakanaAILabs]

Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API.

Our ‘Fugu Ultra’ model matches the performance of Fable and Mythos, delivering frontier capability without the risk of export controls.

Try it: https://t.co/hhO6qTawgb 🐡

[하치@ハチ @hachi_sadam]

GINTANくん(@FurryGintan)の名刺をデザインしました✨
これからの活動も応援してる！