6 ms·
It does change the output, they never said it did not. They said it would not _noticeably_ affect performance.
by unglaublich 16d ago
It does change the output, they never said it did not. They said it would not _noticeably_ affect performance.
- sroussey 16d agoWhich is a lie. Or vacuous statement as Claude might say these days.
- anthonyrstevens 15d agoHow is it a lie?
- sroussey 15d agoThe performance of the output is poor and I can detect the watermark but only in a domain i am in the middle of, if it writes a summary of an email i can not tell, but in claude code in my codebase i definitely can tell. I am complaining about it!
- sroussey 14d agoOr when it tries a different language to explain a word: "The four items the ledger had标 standing open"!
- Smaug123 16d agoIt doesn't necessarily change the output distribution; it depends exactly how it's implemented, and Anthropic haven't told us that. Google's original SynthID paper describes how you can do this. Toy proof-of-concept: Anthropic owns a secret key which is a coin-flip Bernoulli random variable K with p=1/2. You are paying Anthropic to give you X, a Bernoulli random variable with p=1/2. Anthropic changes from their old strategy, "draw from K, then throw it away and flip a coin, each time you ask for a sample", to their new strategy, "draw from K and send it to you". You cannot observe the difference, but Anthropic knows K and so they know when you are repeating its outputs. (Obviously this is a toy example; in reality the distribution is vastly more complicated than Bernoulli, and Anthropic isn't just storing some model outputs to use as K but instead is computing a correlation with a known pseudorandomness source.)
- unglaublich 15d agoBut at low, or zero temperature, the distribution effectively collapses to top-token. This will not be compatible with watermarking. Would explain why Antrophic removed T from the API.
- graboy 15d agoYou have a misunderstanding. Watermarking does not bias the responses in any way. How is this possible? Before: "He leaped at the chance" - 33%. "Jumped at the opportunity" - 66%. After: "He leaped at the chance" - 33%. "Jumped at the opportunity" - 66%. But if you refresh your response from Anthropic 100 times: Before: "Jumped at the opportunity" He leaped at the chance" "Jumped at the opportunity" After: "He leaped at the chance" "He leaped at the chance" "He leaped at the chance" The second one is detectable as being watermarked. davmre has a good explanation that's more in-depth.
- unglaublich 15d agoI understand, but this only applies at non-zero temperatures.