Hillary Sanders
@meatlearner.bsky.social
Machine-learner, meat-learner, research scientist, AI Safety thinker. Model trainer, skeptical adorer of statistics.
Co-author of: Malware Data Science
Co-author of: Malware Data Science
I was surprised at how clear-cut and blatant it was. I mean, two times in a row, closed fingers, correct angle.
Meanwhile, Musk has recently issued public support for the far-right wing AfD party, often described as anti-semetic / extremist.
www.cnn.com/2024/12/20/m...
That + no apology...
Meanwhile, Musk has recently issued public support for the far-right wing AfD party, often described as anti-semetic / extremist.
www.cnn.com/2024/12/20/m...
That + no apology...
Elon Musk endorses far-right German political party, wading deeper into global politics | CNN Business
Musk, the billionaire Trump ally who is playing a public role in the incoming administration, posted in support Friday of Alternative for Germany, or AfD, after the German government collapsed this we...
www.cnn.com
January 22, 2025 at 12:29 AM
I was surprised at how clear-cut and blatant it was. I mean, two times in a row, closed fingers, correct angle.
Meanwhile, Musk has recently issued public support for the far-right wing AfD party, often described as anti-semetic / extremist.
www.cnn.com/2024/12/20/m...
That + no apology...
Meanwhile, Musk has recently issued public support for the far-right wing AfD party, often described as anti-semetic / extremist.
www.cnn.com/2024/12/20/m...
That + no apology...
Nice! Would love to be added (11 yrs in AI, co-author of Malware Data Science, love them NNs)
January 7, 2025 at 4:46 PM
Nice! Would love to be added (11 yrs in AI, co-author of Malware Data Science, love them NNs)
Am I reading this right? Techniques to make the model safe again had almost no effect on non-small models :o.
December 3, 2024 at 9:46 PM
Am I reading this right? Techniques to make the model safe again had almost no effect on non-small models :o.
A response to X is going to be (usually) written by someone socially, politically near X's author, vs some other random piece of content Y.
It's extremely hard to take out sycophancy out of an LLM, trained the way we train them.
It's extremely hard to take out sycophancy out of an LLM, trained the way we train them.
December 3, 2024 at 9:00 PM
A response to X is going to be (usually) written by someone socially, politically near X's author, vs some other random piece of content Y.
It's extremely hard to take out sycophancy out of an LLM, trained the way we train them.
It's extremely hard to take out sycophancy out of an LLM, trained the way we train them.
Say a model learns strategy x to minimize training loss --> Later, min(test loss) involves strategy y, but the model regardless sticks with strat x (inner misalignment).
Assuming outer misalignment, x can be seen as safer than y.
That being said, the better the model, the less this will happen.
Assuming outer misalignment, x can be seen as safer than y.
That being said, the better the model, the less this will happen.
November 24, 2024 at 5:01 PM
Say a model learns strategy x to minimize training loss --> Later, min(test loss) involves strategy y, but the model regardless sticks with strat x (inner misalignment).
Assuming outer misalignment, x can be seen as safer than y.
That being said, the better the model, the less this will happen.
Assuming outer misalignment, x can be seen as safer than y.
That being said, the better the model, the less this will happen.