Research papers definitely need to be more nuanced with the "zero-shot" language. Originally this term was used to describe out of distribution and out of class instances and in the context of metalearning (if you don't know, see under the history section I left for context). This term has been really bastardized and it makes it difficult to differentiate works now. "Out-of-domain" is a fuzzy concept and I think there are some weird usages where people would call something OOD but wouldn't call a test set OOD. OOD classically doesn't mean something not in training data, but not in the distribution of data your data is a proxy for. Certainly the data here is within distribution as it is using LLMs.<p>> Our approach, Binoculars, is so named as we look at inputs through the lenses of two different language models.<p>How is LLM generated data out of domain of LLMs? Specifically their github demonstrates with Falcon-7B and Falcon-7B-Instruct models. Instruct models are specifically tuned on their own outputs. We can even say the non-instruct models are also "trained on" LLM outputs as you're using the outputs in the calculation of the cost functions, meaning they see that data and are using that information, which is why<p>> Unsurprisingly, LLMs tend to generate text that is unsurprising to an LLM.<p>Because they are trained on cross-entropy which directly related to perplexity. Are detector researchers really trying to use perplexity to detect LM generation? That seems odd since that's dependent on the exact thing LMs are minimizing... It also seems weird because the premise from the paper is that human writing has more "surprise" than that from an LM, but we're instructing LMs to sound more human. Going about detection this way does not sound like it would be a sustainable method (not that LLM detectors are reliable and I think we all know they frequently flag generic or standard text, which of course they do if you're highly dependent on entropy).<p>=== History ===<p>First example I'm aware of is the "one-shot" case from[0] (2000) and abstract says<p>> We suggest that this density over transforms may be shared by many classes, and demonstrate how using this density as “prior knowledge” can be used to develop a classifier based on only a single training example for each class.<p>Which we can think of as taking a model and fine tuning (often now just called training) with a single epoch, relying on the prior knowledge that the model learned that is general to other tasks (such as training on cifar-10 should be a good starting point for classifying lions).<p>Then come [1,2] in 2008. Where [1]'s title is "Importance of Semantic Representation: Dataless Classification" and [2] (from Yoshua Bengio's group) is "Zero-data Learning of New Tasks".<p>[1] trains on Wikipedia and then tests semantic classification on a modified 20 Newsgroup dataset (expanded labels) and Yahoo Answers dataset and is about the generalizability of the embedding mechanism cross domain. Specifically they compared Bag of Words (BoW) to Explicit Semantic Analysis (ESA).<p>I'll just quote for [2]<p>> We tested the ability of the models to perform zero-data generalization by testing the discrimination
ability between two character classes not found in the training set.<p>Part of their experiments includes training on numeric character recognition and testing on alphabetical characters. They also do some low-shot experiments.<p>[0] <a href="https://people.cs.umass.edu/~elm/papers/cvpr2000.pdf" rel="nofollow">https://people.cs.umass.edu/~elm/papers/cvpr2000.pdf</a><p>[1] <a href="https://citeseerx.ist.psu.edu/document?doi=ee0a332b4fc1e82a9999acd6cebceb165dc8645b" rel="nofollow">https://citeseerx.ist.psu.edu/document?doi=ee0a332b4fc1e82a9...</a><p>[2] <a href="https://cdn.aaai.org/AAAI/2008/AAAI08-103.pdf" rel="nofollow">https://cdn.aaai.org/AAAI/2008/AAAI08-103.pdf</a>