With more traffic moving off-web and into LLMs, I got curious about what traces we leave "in the weights". My design partner and I built a site in the past few weeks that checks recognition across frontier and small models. It queries many of them in parallel, clusters the responses, and tells you how strongly they recognize you. Happy to answer any questions here!
I'm glad to hear the teenage drummer I used to see when googling myself has gone pro. He's doing pretty well, too, if these models can be trusted.
Retro bowl (& goal) theme spotted
https://play.google.com/store/apps/details?id=com.newstargam...
https://play.google.com/store/apps/details?id=com.newstargam...
I put in my name, and four boxes popped up -- one for "American mathematician", one for "spelling bee contestant", one for "American poker player", and one for "fitness industry entrepreneur".
In fact, both of the first two are me, but I wonder if Claude Opus 4.8 (the only one that hit both of those two) realizes they're the same person? :P
There's a famous son of a star, his fame comes from drug addiction and felonies. I'm doomed to have this bias on my name forever...
Incredible concept and a very well-crafted site. I scored very low, but then very high with my legal name. It seems DeepSeek knows a lot of arxiv papers (or at least, about the authors).
Apparently, I am "in the weights". Kimi gets it correctly: https://www.intheweights.com/p/joseph-jude
After searching "foo", if you try to search "foobar", it deletes "foo" and searches "bar".
I like the Possible Hallucinations feature. Seems like a feature that could stand on its own. Interested in how you separate those out.
This is fun!
Great design and artwork. How did you generate the portraits?
Have you thought about extending this into some sort of pipeline for AIO?
So despite publishing a lot of fiction and poetry I’m apparently most well known for my contributions to the TeX, LaTeX and typography communities. It also thinks I’m a professional athlete having played professional baseball hockey and basketball.
I’m in the weights!
So… what does this mean for the right to be forgotten?
According to this, LLMs have never scraped Slashdot!
Slashdot can now be a safe space :-)
I knew I shared a name with a former NBA Forward, but I did not know about some of the other well-known figures. Sadly (or thankfully?) I'm not listed anywhere, even in the hallucinations.
Initial reaction was "wow! I guess I have the same name as a Canadian actor!" And then I looked it up and figured out that I do not, in fact, share my name with a Canadian actor. Kimi K2 and GLM both hallucinated the same thing.
Some models kind of got it right, but when asked directly, they say they don’t know me. Curious behind the scenes how the correlation works
I am currently on a long laggy line on a resort island. Your timeout is too short, I’m sure it would work if it didn’t cut itself off.
I am in the hallucinations.
Apparently I share a name with a prominent white nationalist activist. Yikes.
Interesting Mistral sort of knew something about me, both gpt and deepseek produced the same answer more or less. I wonder why xD, only gemini knew my online handle mostly github and rust which is interesting.
everybody is apparently some kind of professional sports player...
The two matches for my name are hallucinated athletes. For now.
I typed my boss' name in and it returned:
"No stable person found"...
I am in the top 3% but I don't know if that means a lot or not much :-/
Cool way to get names associated with IP addresses
I hope this isn't costing you a ton in tokens
Will Smith
American actor and rapper > 984 strength · Top 1%
MLB catcher for Dodgers > 255 strength · Top 25%
MICHAEL RYAN HUNGERFORD MASSACRE PERPETRATOR 204 STRENGTH · TOP 35%
For fucks sake.
They all know me to 68%-88% certainty. “Known for my contributions to open source”, yeah, sure, let’s go with that ;)
My real name (542, top 9%) is quite a bit under my username (692, top 6%).
Deepseek seems to know a lot about me!
If I have a strength of just 488, how can that put me in the top 10%! Anyways, fun idea.
Love the graphics, the 8-bit style of the people's portraits is really well done. Are those AI generated?
Nice retro styling
just to be clear, you have each search running on all those models? Self hosted a lot of them right?
I’m the top one! Interesting to see the hallucinations creeping in across the weaker models.
First response for me was also a hallucinated Scottish soccer player who doesn't exist
I am not in the weights
this is super fun, interesting to see how much these LLMs know things without internet as a knowledge base
well, the lower confidence ones got my pseudonym, the higher confidence ones missed entirely and attributed it to a prominent speedrunning streamer.
My real name was attributed to a non-existent famous midfield footballer
There is a 'hallucinations' section on the page, which suggests that the items above that section are not hallucinated. I highly doubt that.
I am, as far a I know the only person in the world with my name. So I searched for my name. I am none of many things this tool tells me I am, for example a right wing politician, a journalist,l and a researcher on solar fuels.
Surprised to find myself in the top 50%. Like... _really_ surprised.
Apparently gpt 5.5 thinks I’m a metal folk musician, i wish.
Not terrible although I am somewhat insulted that QWEN3 8B hallucinated me as the chimp from Jimmy Neutron and no, MISTRAL 3.2 24B, I don't stream on Twitch.
Oh and KIMI K2 0905 completely hallucinated a real name for me (I don't work on Pygame!)
It has me through my open source contributions, but interestingly it claims the same regardless of what I use as a given name, as long as I keep the (fairly unique) surname. So my whole extended family are open source contributors.