second half of a split name

What this neuron does

This neuron fires on the later piece of a name or word that the tokenizer has chopped into pieces, such as the 'yard' in Rudyard Kipling or 'kin' in Rodkin. Turning it up did not produce a clear theme in the model's writing.

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

   species.⏎In the late 1800s, naturalists named the spider Bagheera kiplingi after a panther in British author Rudyard Kipling's 1894 children's book The Jungle Book.⏎

2. #2 peak

   their behavior.⏎Genuine connections are the essential prerequisite to creating a positive classroom culture that can withstand the force of changing social dynamics.⏎Role Model Kindness⏎Rodkin & Hodges (2003) cite evidence that when teachers are warm

3. #3 peak

   at home where they eat rice, fish, chicken and vegetables flavored with ngapi (a pungent, dry, fermented fish or shrimp paste described by Rudyard Kipling as "fish pickled when it

4. #4 peak

   .1 expresses suicidal ideation. This interpretation is misguided. Toward the end of his speech Hamlet does mention that life can be surrendered by a mere bodkin. Here, Hamlet is expressing the frailty of the human body and the

5. #5 peak

   , because with odd enough file names one can cause⏎diff to generate a patch that is syntactically correct but⏎patches the wrong files. The<|endoftext|>Dante Gabriel Rossetti⏎George Henry and Edward Atkinson Hornel⏎Henri

6. #6 peak

   faster clicks higlighted in the green regions of the spectrogram. To hear the recording, click on the spectrogram.<|endoftext|>Title: Do crocodilians get the flu?: Looking for influenza A in

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

   .⏎The next step is to determine the identifier for the drive you want to make an image of. In the terminal window, type in the command⏎sudo fdisk -l⏎Our 1GB drive is /<|endoftext|>During the period from

2. #2 peak

   ten objects from museum collections in Greater Manchester⏎A compass used in a PoW escape, a railway station sundial and part of the world's first commercial computer: objects that tell a history of

3. #3 peak

   361. He moved his capital to Adrionople (now Edirne), Thrace's capital and the second city of the Byzantine Empire.

4. #4 peak

   for playing music. He also took Indian classical music to outside countries and got many awards like Unesco Award, Unicef’s National Ambassadorship, and Padma Bhushan (“Instrumentalists”). ...⏎Not exactly what<|endoftext|>The Challenger Reports

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.