Commenters here seem a little fixated on the fact that the technique scores poorly. This is true and somewhat problematic but exploring the data it looks to me

Commenters here seem a little fixated on the fact that the technique scores poorly. This is true and somewhat problematic but exploring the data it looks to me like this could be more a problem with a combination of the scoring function and the limits of language rather than the methodology. For example, look at https://openaipublic.blob.core.windows.net/neuron-explainer/neuron-viewer/index.html#/layers/17/neurons/3218 https://openaipublic.blob.core.windows.net/neuron-explainer/... It's described as "expressions of completion or success" with a score of 0.38. But going through the examples, they are very consistently a sort of colloquial expression of "completion/success" with a touch of surprise and maybe challenge. Examples are like: "Nuff said", "voila!", "Mission accomplished", "Game on!", "End of story", "enough said", "nailed it" etc. If they expressed it as a basket of words instead of a sentence, and could come up with words that express it better I'd score it much higher.