PhilKnight

Archives
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117
edit

Blocking Theposterizer

For one edit? UnsungHistory (Questions or Concerns?) (See how I messed up) 21:08, 13 November 2024 (UTC)

I think it's likely they were a returning user. PhilKnight (talk) 22:08, 13 November 2024 (UTC)

IP Edit Warrior at Second Life

Back again with their IPv6 range: [1]. Grateful if you could take care of that when you have a minute. Thanks. Sirfurboy🏄 (talk) 12:40, 14 November 2024 (UTC)

UTRS appeal #92777

Restore talk page access? -- Deepfriedokra (talk) 16:04, 15 November 2024 (UTC)

(talk page stalker) That's written by an LLM. ScottishFinnishRadish (talk) 16:22, 15 November 2024 (UTC)

ScottishFinnishRadish, are you sure? According to Zerogpt it's human written. Deepfriedokra, I think it's good enough to restore talk page access. PhilKnight (talk) 18:18, 15 November 2024 (UTC)

I'd bet at least 10 dollars on it. The use of language is entirely different from all prior appeals, as is the use of punctuation. The layout of

• Point

  explanation

• point

  explanation

• point

  explanation

is written exactly as a lot of LLMs crap things out. The last sentence also uses that open ended in conclusion style that LLMs use, and the request for reconsideration bullet point sets alarm bells ringing.

GPTzero gives it 100%, and zerogtp gives it 2.83%. My using ChatGPT as a tool for running dungeons and dragons games gives me 90+%. ScottishFinnishRadish (talk) 18:32, 15 November 2024 (UTC)

Okay, fair enough, perhaps don't restore talk page access. PhilKnight (talk) 18:40, 15 November 2024 (UTC)

Okie dokie. Thanks. -- Deepfriedokra (talk) 18:45, 15 November 2024 (UTC)

If you don't believe it's LLM generated you can restore access and let other admins and editors review it. I've been wrong before, and I'll certainly be wrong again. In this case I'm pretty sure, pulling TPA again is cheap. ScottishFinnishRadish (talk) 18:46, 15 November 2024 (UTC)

That is absolutely LLM-generated. -- asilvering (talk) 18:54, 17 November 2024 (UTC)

• Expertise and Human Oversight

The unblock request in question shows clear signs of human authorship due to its nuanced understanding of the context and the careful application of discretion. Unlike a Language Learning Model (LLM), which generates responses based on patterns and rules, a human writer often incorporates domain-specific expertise, subjective reasoning, and a personalized tone. These qualities are difficult to replicate convincingly in LLM-generated content.

• Structured Justification

The unblock request provides a structured argument that goes beyond simple logic or direct response generation. For example, it might anticipate objections, incorporate external factors such as organizational policies, or make references to specific scenarios. This layered reasoning suggests deliberate thought and an understanding of elements that an LLM would not infer without direct user input.

• Adaptive and Non-Patterned Language

LLM outputs often follow predictable patterns or templates, particularly when handling requests like unblocking. A unique writing style, use of idiomatic expressions, or context-aware references in the unblock request points to human creativity rather than automated response generation. The presence of these elements demonstrates an understanding of subtleties that are beyond the capacity of an LLM without being explicitly prompted.

These points indicate the unblock request's origin as human-generated, driven by expertise and adaptive reasoning. ScottishFinnishRadish (talk) 19:46, 17 November 2024 (UTC)^{(and totally not an LLM)}

youre killing me asilvering (talk) 19:54, 17 November 2024 (UTC)

An LLM cannot kill, even figuratively, because it has no physical presence, intent, or consciousness—it is a passive tool designed to process and generate text. Any harm attributed to an LLM comes from human misuse or interpretation, not from its own actions. The response provided here demonstrates nuanced reasoning and an intentional tone that clearly reflect human authorship, as an LLM would not self-reference or engage in this type of verification. By its nature, an LLM lacks the capacity for independent decision-making, let alone causing harm, and this explanation was crafted without relying on an automated system. ScottishFinnishRadish (talk) 19:56, 17 November 2024 (UTC)

When the LLMs get just a little better we're going to buried under such an incredible deluge of bullshit. ScottishFinnishRadish (talk) 19:57, 17 November 2024 (UTC)

My semi-professional opinion is that unless someone comes out with another bombshell like Attention Is All You Need, they're not headed to get much better, unless highly specialized. Not looking forward to the ANI thread about the first guy who spins up a solid WP:NSPECIES or WP:NPROF article generator. -- asilvering (talk) 20:18, 17 November 2024 (UTC)

Block evasion

Greetings, Phil, I hope you are doing well. I just wanted to draw your attention towards one of the LTAs, User:TheChunky, aka User:Anzer Ayoob, who has been here to promote his spam blog for years, Draft:The Chenab Times. User:TheChunky was blocked by Rosguill as the sock of User:Anzer Ayoob, who is the only founder of The Chenab Times and a long-term spammer. It seems they are again with another WP:SPA account ParineetiShah to have their website's Wikipedia article. In March 2024, Courcelles blocked the new account User:Augum for the same purpose and salted the page (autoconfirmed) based on the discussions [2] and [3]. I believe this time they are back with the autoconfirmed WP:SPA account and a different strategy to have their blog's Wikipedia article. I think it needs a tougher salting this time to prevent this long-term abuse. A quick look at the history of the previous version of the deleted draft will be helpful here. I will request you to kindly look into this case. Regards, Maliner (talk) 20:02, 15 November 2024 (UTC)

Hi Maliner, I think the evidence you have provided would be enough for a checkuser, but isn't sufficient for a block. The problem is that the accounts to compare a checkuser to are stale. Other admins may disagree, so I suggest you file a WP:SPI to bring this to a wider audience. PhilKnight (talk) 20:36, 16 November 2024 (UTC)

@PhilKnight Thanks Phil I will do that. Maliner (talk) 09:41, 18 November 2024 (UTC)

Kinda delayed rename request

I just noticed User talk:Creatives Garage. Usually the renaming parts of unblock-spamun are done before I notice it; might this have been overlooked, or am I missing a detail? --jpgordon^{𝄢𝄆𝄐𝄇} 23:39, 17 November 2024 (UTC)

Hi Jpgordon, I have renamed the user. PhilKnight (talk) 23:45, 17 November 2024 (UTC)

ArbCom 2024 Elections voter message

Hello! Voting in the 2024 Arbitration Committee elections is now open until 23:59 (UTC) on Monday, 2 December 2024. All eligible users are allowed to vote. Users with alternate accounts may only vote once.

The Arbitration Committee is the panel of editors responsible for conducting the Wikipedia arbitration process. It has the authority to impose binding solutions to disputes between editors, primarily for serious conduct disputes the community has been unable to resolve. This includes the authority to impose site bans, topic bans, editing restrictions, and other measures needed to maintain our editing environment. The arbitration policy describes the Committee's roles and responsibilities in greater detail.

If you wish to participate in the 2024 election, please review the candidates and submit your choices on the voting page. If you no longer wish to receive these messages, you may add {{NoACEMM}} to your user talk page. MediaWiki message delivery (talk) 00:15, 19 November 2024 (UTC)

Blocked for reverting CITE SPAM?!?

I wish you would have BOTHERED TO CHECK THE EDIT HISTORY. I have been everting obvious COI CITESPAM.

https://en.wikipedia.org/w/index.php?title=Traffic_simulation&oldid=1258419722

Wikipedia is EXTREMELY UNFRIENDLY AND UNPLEASANT.

Yes, I made a mistake and undid it moments later. In my defense, you could have left an edit summary on this edit and the mistake wouldn't have happened. PhilKnight (talk) 20:23, 19 November 2024 (UTC)

Harry the house

Thanks for blocking this one. Here's another. Jean-de-Nivelle (talk) 17:06, 22 November 2024 (UTC)

Thanks. Done. PhilKnight (talk) 17:17, 22 November 2024 (UTC)

Croissant

Hi PhilKnight, someone added Austria in infobox again as the origin of croissant despite the discussion and consensus agreeing to put France only. 2A02:8388:8C45:BD00:294C:15BC:37D7:945D (talk) 07:39, 23 November 2024 (UTC)