Viv ✓

@Vtrivedy10 · Philly · joined 26 Feb 2013

applied research lead @LangChain Labs, prev @awscloud, phd cs @templeuniv

15 302Followers

1 811Following

6 401Posts total

427.2KViews on collected posts

Posts recentes

Viv ✓@Vtrivedy105 Sep

imo “data research” is a top 3 best things to try with new frontier models it requires taste + persistence, is hard to fully encode in a skill, and is long horizon verifiable -> is output model better? also if it works then that’s a step towards self improving agents, so good 8.4K views · 75 likes · 5 reposts · 2 replies Open on X →

Viv ✓@Vtrivedy105 Sep

Another day another lesson written by Fable about data filters on pre-training! Fable grabbed @Teknium's OpenHermes-2.5 dataset and showed how you could find issues with control tokens in the first 2000 rows from Airoboros. Super interesting stuff! https://t.co/W9I9G9HxZz 11.2K views · 39 likes · 1 reposts · 6 replies Open on X →

Viv ✓@Vtrivedy105 Sep

a super genius that maniacally & tastefully (very important) verifies every last detail is what helps us eventually get to norm treating agents as autonomous work engines very excited to test over time how much i can “trust” Astra to do hard work autonomously, it’s a very high h 7.2K views · 49 likes · 1 reposts · 4 replies Open on X →

Viv ✓@Vtrivedy103 Sep

as teams explore flavors of fine-tuning like SFT & RL, i think it’s helpful for someone on the team to internalize what behaviors each method “encourages” + how this happens + what type of model it produces behaviors - SFT=mode covering & RL=mode seeking - how that relates to ht 5K views · 82 likes · 7 reposts · 4 replies Open on X →

Viv ✓@Vtrivedy1022 Jul

https://t.co/Eb1h1RFPie 395.3K views · 790 likes · 74 reposts · 21 replies Open on X →

Em comparação com contas do mesmo porte

5 posts dos últimos 90 dias, ao lado da faixa de 10K–100K seguidores. aparece para muita gente, mas poucos desses espectadores reagem.

Mediana de visualizações8 357esta conta924mediana para 10K–100K

Alcance, %54.61%esta conta3.62%mediana para 10K–100K

Engajamento, %0.75%esta conta1.52%mediana para 10K–100K

Métrica	Esta conta	Mediana para 10K–100K	Proporção
Mediana de visualizações por post	8 357	924	9.04×
Alcance (visualizações ÷ seguidores)	54.61%	3.62%	15.1×
Taxa de engajamento	0.75%	1.52%	0.49×

Outras contas desta faixa → Comparar com outra conta → Como estas referências são construídas →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

395.3K22 Jul

5K3 Sep

7.2K5 Sep

11.2K

8.4K

Last 5 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.23%22 Jul

1.85%3 Sep

0.75%5 Sep

0.42%

0.98%

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.

What the audience does

Likes29.3%1 035 in total

Reposts2.5%88 in total

Replies1.0%37 in total

Quotes0.7%25 in total

Bookmarks66.5%2 351 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Posts recentes

Em comparação com contas do mesmo porte

Growth & engagement

Views per post

Engagement rate per post

What the audience does

Contas semelhantes