word after a question starter

What this neuron does

This neuron fires on the word right after a question starter, such as the "does" in "What does" or the "many" in "How many", especially in lists of study questions. Turning it up made the output wander into unrelated text rather than anything clearly about questions.

Written by Claude from this neuron's activations and steering outputs. Confidence: medium. Think it is wrong? Flag it (opens a prefilled issue on GitHub).

Fires most on

The text windows where this neuron's activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

1. #1 peak

   for no longer than one to two days in the refrigerator.⏎What does vinegar do to chicken?⏎What Does Apple Cider Vinegar Do for Chickens

2. #2 peak

   ?⏎(b) In what quantity?⏎(c) At what price?⏎(d) When do they want (it)?⏎(e) What kind of

3. #3 peak

   ⏎Write a code that will follow the equation y = x * (x + x) * x.⏎Mile to Kilometer⏎Given A4 sizes find A3's long side⏎Polygon Interior Angle Sum⏎Bit to Gigabyte⏎Find

4. #4 peak

   sold 830 newspapers. His brother James sold 177 fewer newspapers than Ben.⏎(a) How many newspapers did James sell?⏎(b) How many newspapers did they sell altogether?⏎In early 2-step problems, the textbook asks

5. #5 peak

   a very low rank. When the school superintendent looked at the results, he immediately identified the teacher–her husband had died during the second year of rankings.⏎What does this anecdote illustrate? For starters, that value-added assessments can

6. #6 peak

   could be a resource for information you are looking for, in the same sense that resources for getting water might include a lake, a stream or a well.⏎What does "documenting resources" mean?⏎Documenting a resource means putting all the right

Steering

Push this neuron up or down while the model writes, and watch what changes. The outputs were generated ahead of time, so nothing runs in your browser.

Most negative on

Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.

1. #1 peak

   photographs and a short<|endoftext|>NASA Announces Asteroid Naming Contest for Students⏎October 18, 2012⏎Students worldwide have an opportunity to name an asteroid from which an upcoming NASA mission will return the first samples to Earth.⏎

2. #2 peak

   to balance this equaliser? You can put more⏎than one weight on a hook.⏎What is the greatest number of counters you can place on the grid below without four of them lying at the corners of a square?⏎How many different triangles can

3. #3 peak

   and delayed the release? Provide reasons for your decision.⏎You must first post your answer(s) before you can

4. #4 peak

   Land."⏎And for background on the coal industry that's pushing this bad idea, download our report "The Dirty Truth About Coal: Why Yesterday's Technology Should Not Be Part of Tomorrow's Energy Future."⏎Make Mine Illegal⏎The Coeur d'

Which neurons in earlier layers feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.

Activation histogram

How often this neuron fires at each strength. Almost always it does close to nothing (the tall bars in the middle); the snippets above are the rare moments it fired hard, amber for positive and blue for negative.

Logit lens

The words this neuron nudges the model toward (amber) or away from (blue) when it fires. A hint, not a label: many neurons work through later layers instead.