sankalp
@dejavucoder
llm and shitposting into crafting ai products and evals dm open to talk on ai engg/post-training/llm stuff
คุณอาจชื่นชอบ
you can read my latest blogpost: my experience with claude code after 2 weeks of adventure now - some lore why i started using it - it's several features - my current workflow - must know commands almost like a beginner guide or log if you will sankalp.bearblog.dev/my-claude-code…
model be like:
We're doing all kinds of stuff with these models that the public isn't even thinking of yet.
one of the most interesting evals or attempts to test shape rotating powers of a language model i have seen in a while. so simple yet challenging the models in latent space.
can your llm rotate a shape inside it's head? i found out yes but it's a fucking idiot when it comes to the upper layer... why? non uniform spatial reasoning.... here's an eval to test the internal latent reasoning of your models.
love karpathy sensei's koding koding koding energy
Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,…
read this thread an year ago but i didn't fully understand it today lol. cursor codebase indexing is clever in the sense they separate the semantic search (cloud) from actual code access (done locally). they store your codebase embeddings on cloud with reference meta-data. they…
An underrated part of Cursor is our codebase indexing system. It provides efficient indexing/updating without storing any code on our servers. (1/9)
Introducing nanosft, a clean single file implementation of finetuning for chat style model. Loads gpt2-124M weights on nanogpt and does supervised finetuning using just pytorch. a side project that I made recently for some prep. link below :) qts/rts appericiated
the 'you're absolutely right' problem is similar or worse with sonnet 4.5 and more than annoyance, i find it hard to trust the output when the model says so especially for more subjective tasks
asked a difficult technical question to a friend who got claude merch recently and he said wait a second... let me put on my thinking cap first
cursor for x - aimed for power users / ai augmentor lovable for x - aimed for vibe coder / non technical people this is just how i interpret it
looks interesting
Ever wondered how LLMs evolve from predicting the next token to following your instructions? Post-training 101: A hitchhiker's guide into LLM post-training This is a new guide breaks down the basics of LLM post-training, covering the full journey from pre-training to…
United States เทรนด์
- 1. Bears 90.1K posts
- 2. Jake Moody 13.9K posts
- 3. Happy Birthday Charlie 10.7K posts
- 4. Snell 24.8K posts
- 5. Bills 143K posts
- 6. Caleb 49.5K posts
- 7. Falcons 51.7K posts
- 8. phil 181K posts
- 9. Josh Allen 26.9K posts
- 10. Jayden 23.1K posts
- 11. #BearDown 2,395 posts
- 12. Joji 31.5K posts
- 13. #Worlds2025 18.2K posts
- 14. Swift 291K posts
- 15. Ben Johnson 4,466 posts
- 16. Turang 4,373 posts
- 17. #Dodgers 15.5K posts
- 18. Troy Aikman 6,635 posts
- 19. Roki 6,122 posts
- 20. #BeckyEntertainment 113K posts
คุณอาจชื่นชอบ
-
vijay singh
@dprophecyguy -
TDM (e/λ) (L8 vibe coder 💫)
@cto_junior -
sandrone
@sand7one -
shrihacker
@shrihacker -
shaurya
@shauseth -
fudge
@fuckpoasting -
🎙️Ankush Dharkar ☯️
@ankushdharkar -
Indro
@IndraAdhikary7 -
filterpapi
@filterpapi -
gravito
@Gravito841 -
pragun
@pragdua -
Out Of Office⭕
@Parikshit_K_ -
OTAKU
@OtakuProcess -
Nirant
@NirantK -
ankit
@ankitiscracked
Something went wrong.
Something went wrong.