unclear: direction and ordering words

What this neuron does

This neuron is hard to pin down. It fires on words like 'right' in 'from left to right' and on a few other phrase-level words, but the pattern is loose. Turning it up made the model repeat phrases.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. Think it is wrong? Flag it (opens an email to the maker with the neuron already filled in).

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

   center of tourism, but also introduced John Ford to the valley when the director was scouting locations for Stagecoach. The big three are pictured below. From left to right these are the West Mitten, East Mitten and Merrick’s Butte

2. #2 peak

   services and evaluate video quality in six categories of digital defects on a scale of one to five (one being unwatchable and five being spectacular).”⏎“In addition, Comcast is working to

3. #3 peak

   0 miles.⏎Walker is one of the feature personalities in the Pink Palace Museum's African American Trail exhibit.<|endoftext|>Applied Parameters: Index=n (where n = 0-9)⏎This command is used to move the text cursor to a predefined location

4. #4 peak

   ://ocw.mit.edu (Accessed). License: Creative Commons BY-NC-SA<|endoftext|>Make your own D-net. The net is named for the "D" shape of its frame. It⏎is used to collect stream insects from the bottom

5. #5 peak

   back only to 2000 because price indexes for some of the twenty-three items are not available before then.) Bars of the same color are ordered from left to right from the smallest to<|endoftext|>- Development

6. #6 peak

   O P S.⏎No direct anagrams for endoplasm found in this word list.⏎Words formed by adding one letter before or after endoplasm (in bold), or to adelmnops in any order:⏎s - endoplasms

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

   previously-mentioned Three-Year Average of popular species.⏎The Density Rating is the number of fish that were high-quality size or larger sampled per unit of sampling effort. High-quality size, listed in parentheses at the top of the Density Rating column, is

2. #2 peak

   type and material designation. This is designated by the "1N" in the part number. The "1" identifies the component as a diode. The "N"

3. #3 peak

   have a noise reduction coefficient (NRC) of 0.70 or greater. (The NRC indicates the average percentage of sound a material absorbs; an NRC of 0.70 means the material absorbs about 70

4. #4 peak

   , crude death rate, people per physician, percentage of the population under 15 years of age, and per capita Gross National Product (GNP). The higher the rating, the healthier a country's population. For example, the United States

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.