ngrislain

7 ms·

31.
A Practical Method for Testing Differential Privacy – By Andi Cuko(medium.com)
1 points by ngrislain 2y ago|0 comments
32.
by ngrislain 2y ago
Yes it is true that the model has undergone SFT, and RLHF, and other alignment procedures, and hence the logprobs do not reflect the probability of the next token as in the pre-training corpus. Nevertheless, in concrete applications such as
33.
by ngrislain 2y ago
We need to build a syntax tree and be able to map each value (number, boolean, string) to a range of character and then to a GPT token (for which OpenAi produces logprobs). This is the reason we use Lark.
34.
by ngrislain 2y ago
Same token usage. Actually OpenAI returns the logprob of each token conditional on the previous ones with the option logprobs=true. This lib simply parses the output json string with `lark` into an AST with value nodes. The value nodes ar
35.
by ngrislain 2y ago
Actually, OpenAI provides Pydantic support for structured output (see client.beta.chat.completions.parse in https://platform.openai.com/docs/guides/structured-outputs ). The library is compatible with that but does
36.
by ngrislain 2y ago
Haha, I didn't know that one! It's consistent with OpenAI's conception of a "random" dice roll :-D. Joke appart, I'm quite convinced many people would not find 1 or 6 to look "random" enough to be cho
37.
by ngrislain 2y ago
Thank you! Yes indeed, structured output was instrumental in reliably extracting structured data from images from a client.
38.
by ngrislain 2y ago
More precisely it represents the likelihood of seeing this value conditional on the tokens before it.
39.
by ngrislain 2y ago
Thank you! The number is the the sum of the logprobs from the token constituting the individual values. So it does represent the likelihood of seeing this value. So yes OpenAI is super-biased as a random number generator. We sampled other v
40.
Show HN: Value likelihoods for OpenAI structured output(arena-ai.github.io)
115 points by ngrislain 2y ago|42 comments
41.
RAG with Differential Privacy(arxiv.org)
2 points by ngrislain 2y ago|0 comments
42.
DP-RAG(medium.com)
2 points by ngrislain 2y ago|0 comments
43.
Show HN: A simple implementation of Differentially Private RAG(github.com)
2 points by ngrislain 2y ago|0 comments
44.
Nostr – Wikipédia(fr.wikipedia.org)
2 points by ngrislain 2y ago|1 comments
45.
Enhancing May App Service Quality While Safeguarding Patients' Data(sarus.tech)
2 points by ngrislain 2y ago|0 comments
46.
Qrlew SQL Framework(qrlew.github.io)
1 points by ngrislain 2y ago|0 comments
47.
Lake Nyos Disaster(en.wikipedia.org)
2 points by ngrislain 2y ago|0 comments
48.
Anonymization: The imperfect science of using data while preserving privacy(science.org)
3 points by ngrislain 2y ago|1 comments
49.
Discovering New Knowledge While Protecting Privacy(sarus.tech)
1 points by ngrislain 2y ago|0 comments
50.
Compiler-Driven Development in Rust [video](youtube.com)
2 points by ngrislain 2y ago|0 comments
51.
Beyond Few-Shot Learning: LLMs Excel in Synthetic Data Gen with Fine-Tuning(medium.com)
3 points by ngrislain 2y ago|0 comments
52.
Quickly Generate Time-Series Synthetic Data with OpenAI's Fine-Tuning API(sarus.tech)
1 points by ngrislain 2y ago|0 comments
53.
Private Synthetic Data for Generative AI(microsoft.com)
1 points by ngrislain 2y ago|0 comments
54.
Pythagorean cups force their users to fill them in moderation(en.wikipedia.org)
2 points by ngrislain 2y ago|0 comments
55.
by ngrislain 3y ago
Imagine you want to open an access to your DB to someone - Alice - you do not particularly trust. And you want to make sure Alice cannot learn anything about an individual in the database. Maybe you will filter the queries that you assume s
56.
by ngrislain 3y ago
I wrote a minimal demo here: https://github.com/Qrlew/docs/blob/main/tutorials/minimal.ip... There is an interactive playground there as well: https://qrlew.github.io/dp You can cha
57.
by ngrislain 3y ago
Yes Qrlew is based on a research paper ( https://arxiv.org/pdf/2401.06273.pdf ) presented at a AAAI 2024 workshop: https://ppai-workshop.github.io/ As you may know, Differential Privacy is hard to implem
58.
Show HN: Qrlew, simple SQL to SQL-with-privacy written in Rust(github.com)
15 points by ngrislain 3y ago|6 comments
59.
Rewrite your SQL to SQL-with-privacy using Qrlew(qrlew.github.io)
7 points by ngrislain 3y ago|2 comments
60.
by ngrislain 3y ago
Qrlew is an open source library written in Rust aimed at bringing Differential Privacy to SQL

More ›