'that' after say, think and be sure

What this neuron does

This neuron lights up on the word 'that' when it introduces a clause after words like 'says', 'I think' or 'I'm sure'. Its negative side is on crude slang and swear words. Turning it up made the replies wander without that theme.

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

   orer is not in Madison yet, but it's only a matter of time before it arrives here. We have a plan to deal with it."⏎Eddy says that plan was developed

2. #2 peak

   The area of the iceberg spreads over 2,200 square miles, and the volume is 277 cubic miles. How large do you think that is?⏎You might be able to understand it better if the size is compared to

3. #3 peak

   their meeting last month in Hanoi, Vietnam. “It’s actually not a bad idea,” Trump said. Kim answered through an interpreter, saying “I think that is something that could be welcomed.”⏎Sweden permitted a reporter

4. #4 peak

   oak in mid-Missouri survived the Civil War and floods, but this summer's drought has taken a toll on the tree.⏎Landowner John Sam Williamson said that for the first time he can remember he had to haul water into the Missouri River

5. #5 peak

   awesome) discovery. Without knowing too much about it, I’m sure that there will be a raft of problems involving the industrialization of a process like this

6. #6 peak

   Van Honk speculates that the effect does not occur in cautious individuals, because the hormone would make them so paranoid that they would become socially disabled.⏎“I think that people are going to see that testosterone has beneficial effects on social behaviors and carries properties

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

   , T-dick, cock, dick, front hole, vag, boypussy, clitdick, balls, testes⏎- Back hole, arse, butt, bussy⏎- Breasts, chest, breast tissue, boobs

2. #2 peak

   that the US has to send humans to Mars by 2030.” Space is about to get crowded, which means it’s even more important not to screw it up.⏎So Alexander’s panel recommends a few changes. Moving the Office of

3. #3 peak

   entious: "Anything audacious and ambitious has a tincture of pretentiousness in it. Any less than arrogant composer must sometimes feel, ‘Who the hell am I to set forth this

4. #4 peak

   I). However, it wasn’t until 1775 that Alexander Cummings added the u-bend and then 1858 and the Great Stink in London when things really took off. In response to the foul sewage problems of

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.