\
  The most prestigious law school admissions discussion board in the world.
BackRefresh Options Airdrop

Fine-tuning a model on the xo huggingface corpus tp

...
Charismatic flesh therapy
  09/28/26
1.4 million seems like way too few posts compared to the act...
Heady chartreuse french chef business firm
  09/28/26
yeah good point someone needs to scrape more
Charismatic flesh therapy
  09/28/26
"Toxic Data: AutoAdmit is an unmoderated, highly anonym...
Charismatic flesh therapy
  09/28/26


Poast new message in this thread



Reply

Date: September 28th, 2026 6:50 AM
Author: Charismatic flesh therapy



(http://www.autoadmit.com/thread.php?thread_id=5907412&forum_id=2#50158227)



Reply

Date: September 28th, 2026 6:59 AM
Author: Heady chartreuse french chef business firm

1.4 million seems like way too few posts compared to the actual total. what can you even get from that? you're missing so much.

(http://www.autoadmit.com/thread.php?thread_id=5907412&forum_id=2#50158230)



Reply

Date: September 28th, 2026 7:01 AM
Author: Charismatic flesh therapy

yeah good point someone needs to scrape more

(http://www.autoadmit.com/thread.php?thread_id=5907412&forum_id=2#50158231)



Reply

Date: September 28th, 2026 7:03 AM
Author: Charismatic flesh therapy

"Toxic Data: AutoAdmit is an unmoderated, highly anonymous law school discussion board notorious for hosting severe hate speech, doxxing, harassment, and toxic content. If you train a model on this corpus without extreme filtering, your model will heavily inherit these biases, likely generating highly toxic, offensive, or unsafe text."

(http://www.autoadmit.com/thread.php?thread_id=5907412&forum_id=2#50158233)