retr0reg

4 ms·

1.
Rolling in the Diffs(pwno.io)
2 points by retr0reg 2mo ago|0 comments
2.
Heap Overflow in FFmpeg EXIF(bugs.pwno.io)
85 points by retr0reg 9mo ago|34 comments
3.
Death of Gareth Williams(en.wikipedia.org)
1 points by retr0reg 1y ago|0 comments
4.
by retr0reg 1y ago
I work in a ML security R&D startup called Pwno, we been working on specifically putting LLMs into memory security for the past year, we've spoken at Black Hat, and we worked with GGML (llama.cpp) on providing a continuous memory s
5.
Electron Math: I Found RCE in 8M User Note App(retr0.blog)
2 points by retr0reg 1y ago|0 comments
6.
by retr0reg 1y ago
This high dimensional vector space does allows them to theoretically embed a interpretable "deeper understanding", reference https://www.youtube.com/watch?v=wjZofJX0v4M&t=624s . It's the pre-train process
7.
Words are lossy compressor, LMs speaks in meanings(ruikai.posthaven.com)
3 points by retr0reg 1y ago|2 comments
8.
by retr0reg 1y ago
Really interesting perspectives on AIs, it frustrates me how I never really thought of these collateral effects on the rapid innovation of LLMs/AI. It's refreshing to see these counternarrative perspective.
9.
We Found a Heap Overflow in Llama.cpp's Tokenizer(pwno.io)
9 points by retr0reg 1y ago|0 comments
10.
AutoGDB: exploit a binary while sipping coffee(autogdb.io)
3 points by retr0reg 2y ago|1 comments
11.
by retr0reg 2y ago
In fact the llama.cpp codebase is well-developed and actively maintained. It has undergone iterative security hardening, intensive low-level security checks have been implemented in both the core inference engine and RPC components. This st
12.
by retr0reg 2y ago
it can be both because of the unsuccessful leak / wrong `libggml-base` offset. We're building a fake `ggml_backend_buffer` table from the leaked base + offset (the hard-coded offset of `libggml-base` should be adjusted with the co
13.
Heap-overflowing Llama.cpp to RCE(retr0.blog)
248 points by retr0reg 2y ago|54 comments