whoami_nr

6 ms·

31.
Random Thoughts on Investing in AI(rnikhil.com)
2 points by whoami_nr 2y ago|0 comments
32.
by whoami_nr 2y ago
Author here. I just messed up while posting.
33.
by whoami_nr 2y ago
Yes, I am not American and I had no clue about the connotations.
34.
by whoami_nr 2y ago
The Llada paper: https://ml-gsai.github.io/LLaDA-demo/ here implied strong bidirectional reasoning capabilities and improved performance on reversal tasks (where the model needs to reason backwards). I made a logical l
35.
by whoami_nr 2y ago
yeah but you can backtrack your thinking. You also have a mind voice to plan out the next couple words/reflect/self correct before uttering them.
36.
by whoami_nr 2y ago
So, in practice there are some limitations here. Chat interfaces force you to feed the entire context to the model everytime you ping it. Even multi step tool calls have a similar thing going. So, yeah we may effectively turn all of this e
37.
Why I find diffusion models interesting?(rnikhil.com)
202 points by whoami_nr 2y ago|86 comments
38.
by whoami_nr 2y ago
Github pages + Jekyll for the main blog. All the subdomains(mostly side projects which I have put the effort to deploy) are hosted on Replit. Blog: rnikhil.com
39.
LLM Consortium – Query multiple models at the same time(llm-consortium.rnikhil.com)
2 points by whoami_nr 2y ago|1 comments
40.
by whoami_nr 2y ago
Send your prompt to multiple LLMs and get the best response. Based on this tweet from Karpathy: https://x.com/xundecidability/status/1870678482700923325
41.
How to make consistent cold coffees?(rnikhil.com)
1 points by whoami_nr 2y ago|0 comments
42.
by whoami_nr 2y ago
I was learning to code again, and I built this backroom simulator( https://simulator.rnikhil.com/ ) which you can use to simulate conversations between different LLMs(optionally give a character to each LLM too). I think its
43.
You can't bet on everything (and that's okay)(rnikhil.com)
3 points by whoami_nr 2y ago|0 comments
44.
AI Agent Backroom Simulator(simulator.rnikhil.com)
2 points by whoami_nr 2y ago|1 comments
45.
by whoami_nr 2y ago
Made a simple llm backroom simulator. Give the AI agents a name and personality and then watch them get lost talking to each other. Its a lot of fun. you can setup rap battles between random two people, make gandalf and terminator debate th
46.
Browser Based Habit Tracker(habits.rnikhil.com)
2 points by whoami_nr 2y ago|1 comments
47.
by whoami_nr 2y ago
Built entirely using replit agents and hosted on replit. PS: I know this is a clone of many such habit tracking tools. I am re-learning to code again
48.
by whoami_nr 2y ago
Wouldn't 64 in base64 be represented as = ? (and not 10). Love the comic!
49.
by whoami_nr 2y ago
Dynamo AI | ML Application security customer facing engineer, and other ML research engineer roles too | REMOTE/US/Europe Dynamo AI builds evaluation suites for your LLMs to detect hallucination, security and compliance risks. We
50.
by whoami_nr 2y ago
Building LLM evaluation suites. Basically trying to test LLMs for privacy problems (data leakage/memorisation, PII extraction sort of thing), hallucination(RAG, summarisation etc) and security/compliance stuff(like bias/fiarn
51.
by whoami_nr 2y ago
He mentioned it on the Dwarkesh podcast: https://www.youtube.com/watch?v=bc6uFV9CJGg
52.
by whoami_nr 2y ago
Oh my god. This brings back a ton of memories. I was writing a port knocking implementation back in 2016 or so as a side project and I was using this exact flowchart to use iptables for opening/closing ports and routing packets.
53.
Testing LLMs for sequence extraction attacks(dynamo.ai)
1 points by whoami_nr 2y ago|0 comments
54.
How to (accurately) evaluate RAG systems on tabular data(dynamo.ai)
6 points by whoami_nr 2y ago|0 comments
55.
DynamoAI vs RAGAS Comparison(dynamo.ai)
1 points by whoami_nr 2y ago|0 comments
56.
by whoami_nr 3y ago
Thats what the whole thing is about. He is complaining that they don't respect robots.txt
57.
by whoami_nr 3y ago
That’s just jailbreaking(like DAN prompts) and a simpler terminology solution is to stop classifying jailbreaks under prompt injection.
58.
by whoami_nr 3y ago
You can just email me at contact@rnikhil.com. For good measure, I added it to my HN profile too. Not sure about the Cloudfare thing but I just got an alert that bot traffic has spiked by 95% so maybe they are resorting to captcha checks. On
59.
by whoami_nr 3y ago
Thanks. Your blog has been my goto for the LLM work you have been doing and really liked the data exfilration stuff you did using their plugins. Took longer than expected for that to be patched.
60.
by whoami_nr 3y ago
Fair, I agree and shall correct it. I've always seen jailbreaking as a subset of prompt injection and sort of mixed up the explanation it up in my post. In my understanding, jailbreaking involves bypassing safety/moderation featur

More ›