goodside

9 ms·

61.
by goodside 4y ago
Sure — that’s very doable. I don’t have a summarization demo off-hand but it’s well-explored territory.
62.
by goodside 4y ago
It’s not that implausible. It’s trained on many examples of instructions followed by answers, and it’s meant to (and does) generalize to unseen instructions. After enough training, it also generalized to instructions of previously unseen le
63.
by goodside 4y ago
Yes, I noticed this after I posted. Small errors like this become common when instructions reach this length. It randomly forgets to do steps that aren’t written down — it never leaves things blank, but it forgets pieces of compound directi
64.
by goodside 4y ago
Yes. Based on conversations I’ve had with OpenAI staff, Davinci started unexpectedly developing the ability to answer longer questions as they scaled up normal InstructGPT fine-tuning some time in the past year. They don’t take down old mod
65.
by goodside 4y ago
Be sure to read the thread, in particular: https://twitter.com/goodside/status/1557926101615366144?s=21... > A caveat to all of these: I use GPT-3 a lot, so I know the “golden path” of tasks it can do reliably.
66.
by goodside 4y ago
Thanks. I appreciate not everyone likes or is willing to use Twitter, but I’ve yet to find a more convenient or accessible channel for this content. I could start a proper blog and worry about my own prompt formatting, but that complicates
67.
by goodside 4y ago
This definitely isn’t standard prompting — as far as I know, nobody was aware of this technique until I found it this past week. Its precedent is the “format trick” that Boris Power (at OpenAI) showed me, which is essentially my “instructio
68.
by goodside 4y ago
On the contrary, doing a two-stage generation where the second stage simply judges whether a generation is correct can help a lot. It works even better if you give it several generations and let it choose whichever is the most truthful. I w
69.
by goodside 4y ago
Yes. You can, with effort, condition it to respond sensibly with phrases like “I’m sorry, I don’t know how to reverse strings,” or “I’m sorry, I can’t do any math calculation that a human couldn’t do in their head.” But in doing so you dama
70.
by goodside 4y ago
Thanks! Your blog post on using GPT-3 dialog to explain SQL queries was a big inspiration for me to start posting my prompts publicly: https://simonwillison.net/2022/Jul/9/gpt-3-explain-code/
71.
by goodside 4y ago
The Insert API is much less powerful, because you can infill only a single location and you’re limited to communicating the infill content purely through context, without any instruction. The Edit API is more directly adaptable to this, and
72.
by goodside 4y ago
Many! In this example, my question was explicitly, “How many diverse tasks can I stack into a single generation before it becomes unreliable?” If you scroll down in the thread, I explain that these questions are on the “golden path” of task
73.
by goodside 4y ago
In general, anything that has a “textbook” solution is easy. What it’s doing here is more recitation than synthesis. Where it becomes harder, and where my method is necessary, is when you need to specify the structure of the solution yourse
74.
by goodside 4y ago
No. The method relies heavily on the peculiar fine-tuning of the InstructGPT line of models, which are trained specifically to follow MTurk-style prose instructions. I imagine achieving similar results using a non-InstructGPT model would be
75.
by goodside 4y ago
I include OpenAI Playground links for all but the first several of these, which capture not only the exact prompt but the settings used in the generation. I don’t use gists because you need multiple, non-contiguous highlighted spans of text
76.
by goodside 4y ago
The OpenAI API. I’m using text-davinci-002 (the default) for all of these, with temperature=0 for reproducibility/quality.
77.
by goodside 4y ago
Nice! In general these are better if you run them at the lowest possible temperature. I.e., try temp=0 first for deterministic output and then raise slowly if you need to cherry-pick a better generation.
78.
by goodside 4y ago
Clickable version of links: Python, CSV, NDJSON, R, Markdown, and HTML examples: https://twitter.com/goodside/status/1559801520773898240?s=21... More creative, non-program output in HTML and Markdown: https:&#x2F
79.
Tell HN: A new way to use GPT-3 to generate code (and everything else)
285 points by goodside 4y ago|83 comments
80.
A novel method of complete-program synthesis with GPT-3(twitter.com)
2 points by goodside 4y ago|1 comments
81.
by goodside 4y ago
This is a cool demo, but note the author doesn’t actually show the rewritten headlines are better in any way but spot-checking. The conclusion here isn’t that GPT-3 can optimize titles on its own, but that it generates ideas that, when revi
82.
ASCII Art with GPT-3(twitter.com)
1 points by goodside 4y ago|0 comments
83.
GPT-3 following 2000 chars of instructions(twitter.com)
6 points by goodside 4y ago|0 comments
84.
GPT-3 has no idea what letters look like(twitter.com)
1 points by goodside 4y ago|0 comments
85.
by goodside 4y ago
For anyone wondering why the title of the short story is “Lena”, see: https://en.m.wikipedia.org/wiki/Lenna
86.
by goodside 4y ago
Everything green is generated. The “Assumptions:” bit is generated because I wrote this query incrementally. I had it produce the assumptions for the egg question just so I didn’t have to re-type the syntax, and backspaced its (incorrect) o
87.
“You are GPT-3,” a long-form prompt designed to suppress hallucinations(twitter.com)
8 points by goodside 4y ago|2 comments
88.
by goodside 4y ago
Don’t recall specifically. There maybe may have been randomized A/B tests on the “special blend” at some point, I think — it was never spelled out on the site, but I think that was the experimental mix du jour and we tended to use that
89.
GPT-3 prompting: The “hash trick”(twitter.com)
1 points by goodside 4y ago|0 comments
90.
by goodside 4y ago
I left in 2015, as soon as it became apparent the party was over. OkCupid went downhill for a lot of reasons, but overly aggressive A/B testing wasn't one of them.

More ›