Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: Steerling-8B, a language model that can explain any token it generates

guidelabs.ai

Post reply on HN
Print view View original

91–100 of 104 posts

Re: Show HN: Steerling-8B, a language model that can explain any token it generates

#91

Post by Macuyiko »

killerstorm wrote: post #90

This seems to be too coarse-grained to be useful: all sciency content will be "analytical" and associate with sources like ArXiv. But there might be bad, malicious articles on ArXiv, so it doesn't really say anything about veracity. Perhaps this might help to detect some problems like prompt injection - but then it might be more interesting to see those examples.

The input attribution part is interesting though, but I do wonder to which extent that is just assigning some sort of SHAP values to the input tokens, in which case it should be pretty portable to any kind of model.

Re: Show HN: Steerling-8B, a language model that can explain any token it generates

#94

Post by jbonatakis »

jacquesm wrote: post #89
adebayoj wrote: post #56

Earlier quoted context omitted.

You got it exactly right :) And you can update the attribution.md to have it NOT rely on opensource projects that have been compromised. Imagine asking claude code to write a package/function in the style of a codebase that you care about or force it to ALWAYS rely on some internal packages that you care about. The possibilities are endless when you insert such knobs into models.

I would rather see that it does not rely on open source projects that have not given permission to be used to train that particular AI on.

Doesn’t the nature of most open source licenses allow for AI training though?

Example — MIT:

> Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions

Re: Show HN: Steerling-8B, a language model that can explain any token it generates

#95

Post by Imustaskforhelp »

jbonatakis wrote: post #94
jacquesm wrote: post #89

Earlier quoted context omitted.

I would rather see that it does not rely on open source projects that have not given permission to be used to train that particular AI on.

Doesn’t the nature of most open source licenses allow for AI training though? Example — MIT: > Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Softw…

I remember seeing some new licenses like Human license or something iirc but they all had the valid criticism that it would be unenforcable or hard to catch decision.

I haven't looked at the project that much but this could seem exciting to me if maybe these two things can get merged.

I don't think that license is necessarily the problem in here. Licenses can change and people can adopt new licenses.

Re: Show HN: Steerling-8B, a language model that can explain any token it generates

#96

Post by ranger_danger »

wasmainiac wrote: post #80
KingOfCoders wrote: post #66

Earlier quoted context omitted.

Not as long as all developers add an ATTRIBUTION.md citing all open source projects they read the source for, all companies they worked for and trained them and all Stack Overflow answers they have used for write the code.

Not everything has to be symmetrical. I’m sure there is a name for that logical fallacy.

I don't consider it a logical fallacy so much as a philosophical debate on art vs theft that exists in both human and AI worlds.

IMO Nothing and nobody starts out original. We need copying to learn, to build a foundation of knowledge and understanding. Everything is a copy of something else (or put another way, art is more like a sum of your influences). The only difference is how much is actually copied, and how obvious it is.

And in the US at least, from a legal perspective, this "how obvious is it" subjective test is often one way that copyright disputes are settled.

For example there have been many cases of similar sounding songs that either did in fact draw an influence from an existing track (whether consciously or not), or were more likely just coincidental... but courts have ruled both ways in such cases, even if they sound extremely similar.

Re: Show HN: Steerling-8B, a language model that can explain any token it generates

#97

Post by mitchbob »

jzig wrote: post #86
msteffen wrote: post #77

In the recent HN thread announcing the new Gemini coding agent ( https://news.ycombinator.com/item?id=47074735 ), a lot of people complained about Gemini’s tendency to do unwanted refactors, not perform requested actions, etc. It made me cautiously optimistic that all of Anthropic’s work on alignment, which they did for AI safety, is actually the cause of Claude code’s comparatively superior utility (and their presen…

What does alignment even mean? What is being aligned and what is it aligning to?

https://en.wikipedia.org/wiki/AI_alignment

Re: Show HN: Steerling-8B, a language model that can explain any token it generates

#99

Post by ottah »

adebayoj wrote: post #18
ottah wrote: post #10

It's a neat party trick, but explainability it's not solution to any AI safety issue I care about. It's a distraction from real problems, which is everything else around the model. The inflexible bureaucratic systems that make it hard to exercise rights and deflect accountability.

op here. Important point, but I disagree. We see explainability/interpretability as a CORE need for AI safety. We believe you can't align/audit/debug/fix a system that you don't understand. Just to give you some answers for what we can do: 1) We can find the training data that is causing a model to output toxic/unwanted text and correct it. 2) We know what high level concepts the model is relying on for any group of…

> wouldn't you like to know why a model is being sycophantic? Or Sandbagging?

Actually, emphatically no. The only thing I care about is that I have recourse. It shouldn't matter the reason, in fact explainability can be an impediment to accountability. It's just another plausible barrier to a remedy that a bureaucracy can use deny changing a decision.

Re: Show HN: Steerling-8B, a language model that can explain any token it generates

#100

Post by audunw »

The one big thing missing from LLMs is the ability to express how confident it is in the truth of what it’s saying.

Perhaps this could be a step in that direction. If we can associate the attribution with likelihood of being true. E.g., Arxiv would be better than science fiction in that context. But what is the attribution if it hallucinates a citation? Im guessing it would still be attributing it to scientific sources. So it does nothing to fix the most damaging instances of hallucination?

Post reply on HN