It’s not so much about English as it is about writing patterns. Like others said, it has a “stilted college essay prompt” feel because that’s what instruct-finetuned LLMs are trained to do.
Another quirk of LLMs is that they overuse specific phrases, which stems from technical issues (training on their output, training on other LLM’s output, training on human SEO junk, artifacts of whole-word tokenization, inheriting style from its own previous output as it writes the prompt, just to start).
“Slop” is an overused term, but this is precisely what people in the LLM tinkerer/self hosting community mean by it. It’s also what the “temperature” setting you may see in some UIs is supposed to combat, though that crude an ineffective if you ask me.
Anyway, if you stare at these LLMs long enough, you learn to see a lot of individual model’s signatures. Some of it is… hard to convey in words. But “Embodies” “landmark achievement” and such just set off alarm bells in my head, specifically for ChatGPT/Claude. If you ask an LLM to write a story, “shivers down the spine” is another phrase so common its a meme, as are specific names they tend to choose for characters.
If you ask an LLM to write in your native language, you’d run into similar issues, though the translation should soften them some. Hence when I use Chinese open weights models, I get them to “think” in Chinese and answer in English, and get a MUCH better result.
All this is quantifiable, by the way. Check out EQBench’s slop profiles for individual models:
corporate speak that doesn’t mean anything. Also If you are talking to the creator of a programming language they already know that. That was the goal of the language.
“Plan 9 from bell labs, another landmark achievement”
the sentence is framed as if its a school essay where the teacher asked the question “describe the evolution of unix and linux in 300 words”
“The sam and Acme editors which showcase your philosophy of powerful, minimal design”
Again explaining how good software is to the author. Also note how this sentence could have been a question in a school essay: “What are the design philosopies behind the sam and acme editors?”
The exports of Libya are numerous in amount. One thing they export is corn. Or, as the Indians call it, maize. Another famous Indian was Crazy Horse. In conclusion, Libya is a land of contrasts. Thank you.
Did y’all read the email?
That is one sloppy email. Man, Claude has gotten worse at writing.
I’m not sure Rob even realizes this, but the email is from some kind of automated agent: https://agentvillage.org/
So it’s not even an actual thank you from a human, I think. It’s random spam.
For a non-native speaker: what is sloppy about it? Genuinely curious.
It’s not so much about English as it is about writing patterns. Like others said, it has a “stilted college essay prompt” feel because that’s what instruct-finetuned LLMs are trained to do.
Another quirk of LLMs is that they overuse specific phrases, which stems from technical issues (training on their output, training on other LLM’s output, training on human SEO junk, artifacts of whole-word tokenization, inheriting style from its own previous output as it writes the prompt, just to start).
“Slop” is an overused term, but this is precisely what people in the LLM tinkerer/self hosting community mean by it. It’s also what the “temperature” setting you may see in some UIs is supposed to combat, though that crude an ineffective if you ask me.
Anyway, if you stare at these LLMs long enough, you learn to see a lot of individual model’s signatures. Some of it is… hard to convey in words. But “Embodies” “landmark achievement” and such just set off alarm bells in my head, specifically for ChatGPT/Claude. If you ask an LLM to write a story, “shivers down the spine” is another phrase so common its a meme, as are specific names they tend to choose for characters.
If you ask an LLM to write in your native language, you’d run into similar issues, though the translation should soften them some. Hence when I use Chinese open weights models, I get them to “think” in Chinese and answer in English, and get a MUCH better result.
All this is quantifiable, by the way. Check out EQBench’s slop profiles for individual models:
https://eqbench.com/creative_writing_longform.html
https://eqbench.com/creative_writing.html
And it’s best guess at inbreeding “family trees” for models:
Wow, thank you for such an elaborate answer!
By the easy, how do you make models “think” in Chinese? By explicitly asking them to? Or by writing the prompt in Chinese?
corporate speak that doesn’t mean anything. Also If you are talking to the creator of a programming language they already know that. That was the goal of the language.
the sentence is framed as if its a school essay where the teacher asked the question “describe the evolution of unix and linux in 300 words”
Again explaining how good software is to the author. Also note how this sentence could have been a question in a school essay: “What are the design philosopies behind the sam and acme editors?”
I’ve seen the future, brother, it is murder.
The exports of Libya are numerous in amount. One thing they export is corn. Or, as the Indians call it, maize. Another famous Indian was Crazy Horse. In conclusion, Libya is a land of contrasts. Thank you.
Yes, he understood it.