connector words in how-to sentences

What this neuron does

This neuron fires on small connecting words in instructions and explanations, such as 'and', 'a' or 'just', placed right before the next part of a sentence. It is hard to pin down more precisely, and turning it up produced repetitive text about time.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

   Baking powder is a combination of sodium bicarbonate and powdered acid (like cream of tartar). Because it already contains an acid, all it needs is moisture and heat added to activate it.⏎Can vinegar and baking soda be mixed?⏎Here are

2. #2 peak

   help our planet.⏎You will need:⏎- Scrap paper⏎- Old books/newspapers/leaflets etc⏎- Needle and thread⏎-

3. #3 peak

   available word search grids cover subjects ranging from animals and timbers via countries and capitals to Shakespeare and mythology.⏎All that is needed for your own word search game is a list of words, or short word groups as spaces and other non-alphabetic

4. #4 peak

   04 op-ed for USA Today, Jonathan Turley, a law professor at George Washington University, argued that polygamy should be legal as a simple matter of equal treatment before the law. Turley acknowledged that underage girls are sometimes forced to

5. #5 peak

   sun. Wear sunglasses, a hat and a shirt, especially in the middle of the day (even on over

6. #6 peak

   IVE (10 minutes)⏎Display the following items on a table:⏎Telephone Newspaper Letters & Envelopes Road Map Radio⏎As the children arrive, give them a sheet of paper and ask them to list what they think the items on the table

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

   +3 and 4y = –2x+10 with y = mx + c⏎The slopes are m = 2 and m = –1/2⏎If the slope of one line is the negative reciprocal of the second line

2. #2 peak

   2. Thus, the probability is about cxn, where c = (10/9)2 ≈ 1.2 and x = (9/10)log102 ≈ 0.97. The

3. #3 peak

   10 log10(MAX^2/MSE) = 10 log10(MAX/(MSE)^(1/2))^2 = 20 log10(MAX/(MSE)^(

4. #4 peak

   10).⏎Article last edited on: 2008-09-09 15:47:35⏎Did you find this article helpful? Yes No⏎How can we improve this article?<|endoftext|>Java tutorials

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.