martianlantern

4 ms·

31.
by martianlantern 1y ago
Do mean hard to be used because of compute required to render it or because of setting up and understanding the training pipeline and related hyperparameters?
32.
Paged Attention Performance Analysis(martianlantern.github.io)
1 points by martianlantern 1y ago|0 comments
33.
Paged Attention Performance Analysis(martianlantern.github.io)
3 points by martianlantern 1y ago|0 comments
34.
by martianlantern 1y ago
Do you know you can achieve the exact same functionality with the vscode's inline code execution in python files and can create jupyter like code cells with '# %%'? ref: https://code.visualstudio.com/docs&#x2F
35.
ThinkMesh: A Python lib for parallel thinking in LLMs(github.com)
73 points by martianlantern 1y ago|4 comments
36.
Train a small GPT to solve mazes(github.com)
1 points by martianlantern 1y ago|0 comments
37.
Reasoning Cot Injection(github.com)
1 points by martianlantern 1y ago|0 comments
38.
by martianlantern 1y ago
Is there any benchmarks and comparisons compared to gpt-oss? I believe it far exceeds gpt oss or even gpt5 otherwise they wounldn't have released it
39.
by martianlantern 1y ago
Hey, really cool work love the idea of focusing on key decision points. I was curious though since confidence can be non monotonic during CoT[1], how does binary search handle cases where there are multiple ups and downs in confidence? It s
40.
Injecting doubts in the CoT of reasoning models(github.com)
2 points by martianlantern 1y ago|0 comments
41.
Injecting doubts in CoT of reasoning models(github.com)
4 points by martianlantern 1y ago|0 comments
42.
by martianlantern 1y ago
There's no explanation as to how they achieved that speed up :( it would have been better if they also wrote a post on that
43.
by martianlantern 1y ago
I’m not familiar with Zig and would appreciate an explanation of how this works. My understanding is that cache behavior is managed by the CPU, and programmers only influence it indirectly through the sequence of instructions (i.e., access
44.
by martianlantern 1y ago
But why?
45.
by martianlantern 1y ago
Am I the only one who is not able to access the interactive demo?
46.
by martianlantern 1y ago
Is there even anything not shortsighted till now?
47.
by martianlantern 1y ago
Seems very handy, is there a way to hide the record pop up so we can concurrently do other tasks?
48.
by martianlantern 1y ago
What types of projects do you use cursor heavily for and what did you stop doing because of cursor, what did you start doing more of?
49.
by martianlantern 1y ago
Very insightful post, this may work in the IMO setting because mathematical problems are inherently binary if we ignore somethings like the incompleteness theorem. In contrast, subjective tasks, such as evaluating a painting or rating a poe
50.
by martianlantern 1y ago
> Laymen used LLMs for all kinds of things as well. In an interview a prominent Finnish actor admitted to (paraphrasing) “pasting a full script into an AI app for content analysis and structuring” I think it is now soon that hollywood mo
51.
Show HN: Kubernetes Tutorial(kubernetes.io)
1 points by martianlantern 1y ago|0 comments