I did a project that attempts to generate these types of instructions: https://map2seq.schumann.pub/nllni/demo/

You can simply pre-render them and store the png tiles in a folder. The folder structure and naming has to be such that e.g. leaflet understands it. Did this for a research demo here: https://map2seq.schumann.pub/nllni/demo/

Is there actually a simple way to pre-render? AFAIK it's a lot of work and learning at least if you're coming from scratch.

QGIS can make a tile pyramid of the current map/project at a specific set of bounds.

I've used TileMill to create tiles for a specific layer of data for a specific area. (Fire evacuation zones for a county.)

https://github.com/tilemill-project

"hallucination" was coined in the context of text generating RNNs. Specifically in this blog post by Karpathy in 2015: https://karpathy.github.io/2015/05/21/rnn-effectiveness/

1-2 seconds in Overpass which also uses OSM:

https://overpass-turbo.eu/s/1NTy

we systematically tested this: https://arxiv.org/pdf/2308.16060.pdf

Where can we use it ? I couldn't find a link. It is closed-source ?

OK, I had a hard time understanding that the link given above is the demo of this paper.

https://overpassnl.schumann.pub.

Very interesting. Quite surprising too : it can find non trivial searches like "bars with darts" ("bars à fléchette" in French), but not easy ones ("fromagerie" or "dermato").

More complex queries did not work : "cafes close to the river" returned nothing, neither did "touristic site close to a pedestrian area" or "touristic streets".

Yes, it's far from perfect. We currently work on augmenting the prompt with e.g. relevant key-value pairs extracted from https://wiki.openstreetmap.org/wiki/Map_features

Tell me more about that coal burning record in Germany.

You can always argue that the model has seen some variation of a given problem. The question is if there are problems that are not a variation of something that already exists. How often do you encounter truly novel problems in your life?

There is a paper?

The grandparent is probably talking about the InstructGPT paper? But I don't remember seeing a preference for longer responses in that paper.

I meant the blog post.

https://openai.com/blog/chatgpt/

> The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. These issues arise from biases in the training data (trainers prefer longer answers that look more comprehensive) and well-known over-optimization issues.12

> Stiennon, Nisan, et al. “Learning to summarize with human feedback.” Advances in Neural Information Processing Systems 33 (2020): 3008-3021. ↩

> Gao, Leo, John Schulman, and Jacob Hilton. “Scaling Laws for Reward Model Overoptimization.” arXiv preprint arXiv:2210.10760 (2022). ↩

I get motion sick hearing brown noise. I associate it with being on a plane or train but I am sitting still, so it feels off and I get dizzy.

I actually bought this as a physical book on Amazon. Naturally it came as a print-on-demand book. Unfortunately it has many problems in this format. E.g. the lack of margins makes it hard to read the end of sentences towards the gutter. Also some text is pushed into each other. Not sure what source file format you have to provide to Amazon, but it's certainly not the pdf provided in the repo.

Edit:

It seems the overlapping text also occurs on some pdf readers: https://github.com/BoltzmannEntropy/interviews.ai/issues/2

The last 5 textbooks I bought new on amazon had similar problems. Totally unacceptable. I started returning them and (because most were exclusive to amazon) started buying them new on ebay with great results.

It's really a hit and miss. This [1] book also came as print-on-demand but looks perfectly fine. Good layout and clean colors.

[1] https://mml-book.github.io/book/mml-book.pdf