ElProthereum (τ,τ)

Sharing daily insights on Al, Tech…

[Trishool | SN23 @trishoolai]

We’re excited to announce that Trishool’s HaloGuard 1.0 𝐡𝐚𝐬 𝐚𝐜𝐡𝐢𝐞𝐯𝐞𝐝 𝐒𝐎𝐓𝐀 prompt-safety performance among open-weight guard models.

Today, we present HaloGuard 1.0, a constitutional input classifier for multilingual AI safety.

It is built as a first-layer input guard that checks user prompts before they reach a downstream LLM, agent, or application.

This is part of the safety infrastructure being built through @trishoolai , our decentralised AI red-teaming subnet on Bittensor SN23.

Full arXiv paper goes live soon.

[const @const_reborn]

@AnthropicAI And just like that we collectively saw the future of inequality

[ElProthereum (τ,τ) @aa_w_lay]

@perplexity_ai @PPLXDevs @grok is it RLM like ?

[WallStreetBets @wallstreetbets]

https://t.co/M1xMCyfU4j

[ElProthereum (τ,τ) @aa_w_lay]

@_mm85 So cool! I’ll be there in a couple of days. Do you mind sharing where this is and any tips for visiting? Thx!

[ElProthereum (τ,τ) @aa_w_lay]

@markjeffrey @DavidSacks @garrytan 👀

[ElProthereum (τ,τ) @aa_w_lay]

@kimmonismus @grok where is this ?

[ElProthereum (τ,τ) @aa_w_lay]

@kimmonismus @grok where is this

[Macrocosmos @MacrocosmosAI]

Today, we present ResBM (https://t.co/EqfVZRerDY), a 128x activation compression technique for achieving SOTA training results in low-bandwidth, distributed communication settings for pipeline parallel training across the internet.

This technology underpins @IOTA_SN9 - our distributed training platform.

[ElProthereum (τ,τ) @aa_w_lay]

❤️👀

[siamkidd @SiamKidd]

Love it when @const_reborn goes beast mode... Still struggling to comprehend how they can do a distributed run of 1T this fast!

[Macrocosmos @MacrocosmosAI]

Training frontier models over the internet requires new techniques.

Today, we present ResBM, a residual encoder-decoder bottleneck architecture that enables 128x activation compression for low-bandwidth distributed pipeline parallel training.

Developed for @IOTA_SN9, we show SOTA compression without significant loss in convergence rates, increases in memory, or compute overhead.

Expect the full paper release in the next 72 hours.

[const @const_reborn]

https://t.co/IdB9DPr5u0

[ElProthereum (τ,τ) @aa_w_lay]

@AlertesInfos @grok verify

[ElProthereum (τ,τ) @aa_w_lay]

Amazing !

[Affine @affine_io]

One team can't build the best AI.

So we stopped trying.

We built an open arena instead.

Here's what happened👇

[Affine @affine_io]

One team can't build the best AI.

So we stopped trying.

We built an open arena instead.

Here's what happened👇

[ElProthereum (τ,τ) @aa_w_lay]

@birdabo @grok verify

[ElProthereum (τ,τ) @aa_w_lay]

@WW3_Monitor @grok what is he talking about

[ElProthereum (τ,τ) @aa_w_lay]

@WW3_Monitor @grok verify

[ElProthereum (τ,τ) @aa_w_lay]

@wesbos @grok explain what he did

[ElProthereum (τ,τ) @aa_w_lay]

@camille_moscow @grok true ?