from the mod log:

Mojng5xTnVM7YHd.png

https://s.faf-pb.xyz/lXxek was the old url but they took it down - I have a copy at https://join.piefed.social/wp-content/uploads/2026/05/evidence-llm-used-for-banning.txt

The output from the LLM was provided as proof that someone needed to be banned.

I didn't want to do this but my hand was forced by their dissembling, minimization and bullshit. Note how the output includes a nice big logo with a slogan and ascii graphics - this was not a one-off experiment, this is an app they were proud to call their own.

you are viewing a single comment's thread
view the rest of the comments
[–] 27 points 6 months ago (4 children)

This is a whole lot of absolutely nothing. Thats why.

BREAKING NEWS! Mod team develops and uses mod tool!

  • source
  • parent
  • hideshow 4 child comments
  • [–] 10 points 6 months ago* (last edited 6 months ago) (3 children)

    „Anarchist” mod team delegating work to a non-deterministic tool operated by big tech is the news here.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 7 points 6 months ago* (last edited 6 months ago) (2 children)

    There is no delegation to the tool.

    It provides an evaluation in text format, and references. Thats it.

    Edit: To be clear, rimu is just making that part up. It has no ability to auto-ban and its not an automod.

    Its a python script to check a user's comment history and evaluate using a local model, and spits out what it finds. A person has to read it and make a decision on this information.

    Everything else is rimu's hallucination.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 4 points 6 months ago* (last edited 6 months ago) (1 child)

    If there was a full review of source text that was summarised by an LLM then it wasn’t needed. It’s a pointless tool that makes guesswork look very confident. This guesswork was done by a black box ran by OpenAI.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 10 points 6 months ago* (last edited 6 months ago)

    If there was a full review of source text that was summarised by an LLM then it wasn’t needed

    You go through every single users comments?

    Do you realize how long that would take?

    I disagree, its a huge time saver.

    This guesswork was done by a black box ran by OpenAI.

    Local qwen, summarized with receipts.

  • source
  • parent