thomadev0

3 ms·

1.
I Let an AI Agent Hack All My Gadgets–and I'd Do It Again(wired.com)
4 points by thomadev0 2mo ago|0 comments
2.
by thomadev0 2mo ago
Based on GLM-5.3, which is #3 on Terminal-Bench 4.0 (behind only Opus 5 and Fable), with 2× the cyber exploitation of 5.2. We abliterated and hosted it so it does the offensive cyber, red teaming, and agent testing work other models refuse
3.
Show HN: Abliterated GLM-5.3 API (84.5% CyberGym, FP8)(abliteration.ai)
2 points by thomadev0 2mo ago|1 comments
4.
Best US- and EU-hosted abliterated models with no input/output prompt retention(next.abliteration.ai)
3 points by thomadev0 3mo ago|2 comments
5.
by thomadev0 3mo ago
We host the best abliterated and fine tuned models with no input or output prompt retention based here in the Bay Area. Integrates with all major harnesses including Claude code, codex, opencode, cyberstrike, promptfoo, Garak and etc… We ha
6.
Show HN: Abliterated GLM 5.2 model, fine-tuned for security and red-team work(abliteration.ai)
2 points by thomadev0 4mo ago|1 comments
7.
by thomadev0 6mo ago
A bit more context on what this is: Abliteration now has a training-data generation workflow. You describe the examples you want, and it generates a dataset for that use case. If the examples need current or real-world facts, there is an op
8.
Show HN: Abliteration – made-to-order training data for classifiers and evals(abliteration.ai)
1 points by thomadev0 6mo ago|1 comments
9.
by thomadev0 1y ago
No model names, evals, or sizes. It would be nice to have more info on the models.