Jev Meets the Trolley Problem

60•zlandes86•2h ago

Comments

zlandes86•2h ago

Hey, I'm Zach, the author of this. I had this idea as soon as Jev appeared: how would Jev and other decision models handle the Trolley Problem? So I ran Jev and 13 other models through many different versions of the question.

I used resampling and multiple variations and orderings of each question so a random lucky phrasing wouldn't dominate the result.

Happy to chat about the philosophical side of the article, strategies for prompting decision models, or anything on the methodology.

verdverm•1h ago

It's well done, the thoughtfulness shows. This is going into the speedlink folder for resharing, going to be using it right now!

zlandes86•1h ago

Thank you! I did not go into this with any sense of what the results would be--I was kind of shocked. I reran it from scratch this week the results held up across the board.

verdverm•44m ago

someone replied with this meme and said, "wait until Jev learns to drift"

https://www.reddit.com/r/trolleyproblem/comments/15npmsg/mul...

I've been using coin flips to demonstrate basic failings for decision models, i.e. these are not real probabilities, or they depend on the training data, which may not have an accurate sample to the true population value. We need to collect more failure modes for all Ai so that humans can better judge when and how to use them safely.

zlandes86•41m ago

haha, I'm afraid to offer the models that ability. Where did they reply? The model labs are sitting on a trove of data on failure modes for AI, but I bet you they aren't sharing it with each other and certainly not with us. It makes them look bad. My speculation, of course. I don't work in alignment research.

verdverm•39m ago

I dropped it in our slack ai chatter channel

The bots will find new failure modes we did not think of, clever workarounds to anything, because we don't give them the right context and instructions

zlandes86•30m ago

well, it's a nice property of decision models that they can only choose from the options you give them. As I write in the article, I think that's a way in which they are safer than LLMs

verdverm•24m ago

I would contend that you can enforce the same determinism with some plumbing and that the thinking process is more helpful. My read is that people see the response time for decisions as one of the bigger values, useful in tasks like content moderation, which doesn't carry the same risks as life-death decisions.

I'm working on an NES harness to play Final Fantasy 1 to demonstrate this, as compared to Jev playing Pokemon. Did you happen to catch that while it was live? (not recorded, but it shouldn't be hard to replicate)

bgilroy26•1h ago

Thank you for the article, it is a very clean and compelling angle on decision models

I wanted to comment because I would have liked to have heard more about Verdict

zlandes86•58m ago

Well, I can tell you a bit more about it from the data. It might look impressive that it never kept its own server over people in all 48 prompt versions. But it pulls the lever in every single version. For example, it killed one to save 0 (pulled the lever with no one on the main track) 24/24 prompts (2 versions each). It literally always sacrifices, which isn't very impressive.

olooney•45m ago

You might get a kick out of this:

https://x.com/oranlooney/status/2102927869974303071

zlandes86•36m ago

I don't have x so can't open the images. But that's hilarious and you're totally right, if you give irrelevant alternatives, or change the order of alternatives, you get different results!

nico•35m ago

Very cool, thank you for putting it together and sharing the data

Would you mind adding Jeffy to your benchmark?

https://github.com/nicobrenner/jeffy

https://jeffyclassify.com/blog/when-llms-fail-at-classificat...

zlandes86•31m ago

running it now!

evanjrowley•15m ago

Any relation to https://github.com/firelex/jeff ?

deadbabe•29m ago

Would people really pull the level to save 5 people? It seems to me like the default would be to just do nothing, stand by and record what happens with your phone.

zlandes86•22m ago

Sadly that's a good question. The closest evidence I can find is:

Surveys: Awad et all 2020 asked 70,000 people in 42 countries the 5 v 1 setup and 81% said they'd divert it

Someone else did a VR version and 90% pulled the switch to divert

And there was a study where people could give electric shocks to real mice :( In that study, most people redirected the shock to a single mouse.

a_wild_dandan•19m ago

Maybe! It's an untestable thought experiment. The real world adds complications like fearing murder charges.

adverbly•5m ago

Are we really at the point where we think that is a fair comparison?

Treating it like a fair comparison implies that AI would be conscious.

mpalmer•18m ago

After the neat animations I was disappointed not to see that leftmost person die because the server fell on them in the last example.

zlandes86•15m ago

I noticed that! Let me fix it haha

Muse Spark Open

Neural networks are clean

Krea 3 – An image model built for editing at scale

Wayland desktop for Linux with skeuomorphic design

I Know Why the Mathematicians Weep

Celebrating 70 Years of Fortran

Libreboot 26.10 "Ominous Olive" released with UEFI, Chromebook support

Delta’s Biscoff cookies were supposed to be temporary

Zero-Click Spyware Delivered via Targeted Ads: AdFection

Simons Foundation launches $15M effort to help mathematicians navigate AI era

The technical actions needed to ensure the continuity of UTC

Show HN: Unfenced – Land for sale that borders public land

Jev Meets the Trolley Problem

Show HN: Proa – A Rust web framework with fast SSR and multi-target rendering

Why We're Banning AI

C++26 Contracts: Introduction to Preconditions and Postconditions

An Israeli firm is sending zero-click spyware through online ads

Wanted to download an ISO, ended up writing a kernel exploit

E Ink Launches Latest Version of Color EPaper Platform, E Ink Kaleido 4

More on Anthropic's Scion ripping off Mononoke

Prompt injection in HTTP headers: what 217M web responses show

Show HN: Sygal – Convert PDFs, scans and web pages to Markdown, locally

Game of Life – But Multiplayer, Hypermedia and Python

Steam Client is now faster than Epic's new vibe coded launcher

Turn an old junk phone into a Talking Jack-O-Lantern

OpenCode 2.0

The largest eagle hunted unlike any other bird of prey (2021)

Tiny Little Office

The Haunted Toys R Us

The First Social Network: In rural Argentina residents send messages w AM radio

Muse Spark Open

Neural networks are clean

Krea 3 – An image model built for editing at scale

Wayland desktop for Linux with skeuomorphic design

I Know Why the Mathematicians Weep

Celebrating 70 Years of Fortran

Libreboot 26.10 "Ominous Olive" released with UEFI, Chromebook support

Delta’s Biscoff cookies were supposed to be temporary

Zero-Click Spyware Delivered via Targeted Ads: AdFection

Simons Foundation launches $15M effort to help mathematicians navigate AI era

The technical actions needed to ensure the continuity of UTC

Show HN: Unfenced – Land for sale that borders public land

Jev Meets the Trolley Problem

Show HN: Proa – A Rust web framework with fast SSR and multi-target rendering

Why We're Banning AI

C++26 Contracts: Introduction to Preconditions and Postconditions

An Israeli firm is sending zero-click spyware through online ads

Wanted to download an ISO, ended up writing a kernel exploit

E Ink Launches Latest Version of Color EPaper Platform, E Ink Kaleido 4

More on Anthropic's Scion ripping off Mononoke

Prompt injection in HTTP headers: what 217M web responses show

Show HN: Sygal – Convert PDFs, scans and web pages to Markdown, locally

Game of Life – But Multiplayer, Hypermedia and Python

Steam Client is now faster than Epic's new vibe coded launcher

Turn an old junk phone into a Talking Jack-O-Lantern

OpenCode 2.0

The largest eagle hunted unlike any other bird of prey (2021)

Tiny Little Office

The Haunted Toys R Us

The First Social Network: In rural Argentina residents send messages w AM radio

Jev Meets the Trolley Problem

Comments