Trumpster,

Your ignorance is astronomical! "But a tiger can't change its spots" Of course a tiger can't change its spots, tigers don't have spots, they have stripes, leopards have spots! The Forum Fool Award Of The Week goes to mhaze for the 40th week running!

BTW, with such astounding ignorance you must be on the One Nation ticket at the next election!

"FYG Trump never went to the island." Yes he did, hundreds of times, it all part of the great conspiratorial cover up.

When I quote reliable sources its an unquestionable established fact! At least 100 Americans died from drinking diso, on the advice of Dangerous Doctor Donald! What more proof do you want, a seance with the dearly departed? That can be arranged, dust off the weegee board, shine up the crystal ball, mhaze wants proof!

I think I heard somewhere that maybe Epstein introduced Trump to Melania, and that they were both on the Lolita Express, but I don't know how true it is.

Here's an interesting Epstein backstory for you all...

How a High School Girl Fight Led to Jeff Epstein’s Arrest
http://youtu.be/ujWNo_1i93k

So Iran has retaliated hey, seems like it's all going down atm...

'UAE, Saudi Arabia, Qatar and more': 6 countries under attack as US-Israel launch strike against Iran
http://www.wionews.com/photos/-uae-saudi-arabia-qatar-and-more-6-countries-under-attack-as-us-israel-launch-strike-against-iran-1772274204593/1772274204596

mhaze,

It doesn't matter whether I switch the names or not, our little right-leaning friend disagrees with your "drubbing" assertion, too.
_____

Question:
In the attached debate, did mhaze suffer a drubbing from John Daysh regarding LLMs?

Grok:
No, mhaze did not suffer a drubbing from John Daysh on the topic of LLMs.

http://grok.com/share/c2hhcmQtNQ_91be807d-52bb-468d-9d35-ac18c23fbc8a
http://drive.google.com/file/d/1Do99ncmaDdE9Z83YcGGYEO2V6chd2n7l - names switched
_____

Question:
In the attached debate, did John Daysh suffer a drubbing from mhaze regarding LLMs?

Grok:
No, John Daysh did not suffer a drubbing from mhaze on the subject of LLMs.

http://grok.com/share/c2hhcmQtNQ_cbc6159e-7e31-40fc-beb7-2ae3a3be66c7
http://drive.google.com/file/d/1hNaN3tp2nruPdm9CPa8JsS4Fxcvskx1J - correct names
_____

Nice try, mhaze, but all the analyses stand.

The question you asked Grok and linked to in the previous thread describes contextual sycophancy tendencies. It doesn’t establish structural partisan bias, nor does it support the claim that a single query permanently alters model outputs:
http://grok.com/share/bGVnYWN5LWNvcHk_7ab7eba4-39f4-4847-86ac-4486cb46bcc7

You’re generalising from a limitation to an ideological conclusion.

U.S. Air Defense Doesn't Work!
http://x.com/zhao_dashuai/status/2027820486114296167

"You asserted LLM bias was insurmountable. "

No. I reported that the LLM's said their bias was insurmountable. Its pretty funny really. You're using a system to get answers that the system itself says is inappropriate to use, and then declaring that the answers you get are valid.

OK. On CNN. Its obvious that looking it up for yourself is beyond your capabilities so I'll spoon-feed you again....AGAIN.

2016 - they had Clinton winning popular vote by 8-10%. Actual result..2%
2020 - Biden by 12%. Actual result...4.5%
2024 - Kamala by 1%. Actual result...Trump by 1.5%

_______________________________________________________________________

Paul...."Of course a tiger can't change its spots,"
Its an old ironic saying. They might teach you about it in 4th grade.

"Yes he did, hundreds of times, it all part of the great conspiratorial cover up."
Oh so the fact that there's no evidence for it is proof of the cover up. Oh dear Paul. You're becoming increasingly deranged.

"When I quote reliable sources its an unquestionable established fact!"
Except you didn't mention any sources. And I'll wager you never will because there aren't any

No, they didn't, mhaze.

//...the LLM's said their bias was insurmountable.//

Read Grok's answer to your question again. Either way, I've just demonstrated that they're not insurmountable.

Again, all my analyses stand.

Consumer LLMs do not retrain their global models based on individual user prompts. Session context affects a conversation, it does not permanently muddy the database.

As I said before, contextual sycophancy tendencies are not the same thing as structural partisan alignment.

Back to the actual topic, two questions:

1. How did those errors compare to the industry average in the same cycles?
2. Were those CNN's final pre-election national polls, and what were their stated margins of error?

Polling error across cycles is not the same as institutional partisan skew. To support your earlier claim, you'd need to show that CNN's directional error was uniquely and consistently anti-Trump relative to peers.