'from' after an excerpt or source noun

What this neuron does

This neuron lights up on the word 'from' when it follows words like 'excerpt', 'scenes' or 'passage' and introduces where something came from. Turning it up made the model's replies drift toward books and passages, though the steered text was loosely related at best.

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

   oval Cherokees by anti-removal Cherokees.<|endoftext|>Create your very own starry night painting inspired by the famous artist Van Gogh using items from around your house,

2. #2 peak

   to allow proper visual development. Ptosis repair is usually completed under general anesthesia in<|endoftext|>BBB - A+ Rating⏎- the best there is⏎An excerpt from www.HouseOfNames.com archives copyright © 2000 -

3. #3 peak

   to evaluate how the policy of coexistence has been implemented in the past ten years.<|endoftext|>BBB - A+ Rating - the best there is⏎An excerpt from www.HouseOfNames.com archives copyright © 2000 -

4. #4 peak

   their goal of developing instrumentation to measure plant photosynthesis and plant characteristics (greenness, leaf area index and biogenic emissions) on its campus.<|endoftext|>Excerpt from the book, Tithes and Offerings⏎Jesus told certain men that if

5. #5 peak

   period (1603-1868). It is a form of woodblock printing that was popular among the common people of Japan, depicting scenes from everyday life, landscapes, and popular culture. Ukiyo-e prints were affordable

6. #6 peak

   “artificial intelligence.” And while it may sound like something out of a science fiction novel, the truth is that AI is very real and it is

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

   rather⏎than singular morphology (and a few languages where there is a split)⏎Swedish<|endoftext|>Old and New London: Volume 2. Originally published by Cassell, Petter & Galpin, London, 1878.⏎This

2. #2 peak

   slan is one of Techfrag's regular news writers. When not digging out news stories, he is found buried nose deep in fantasy novels, wasted listening to dubstep, or raging against Xbox Live kids. Need I say more?<|endoftext|>New

3. #3 peak

   in their September 1915 issue. The use of aerial detection to spot and report forest fires—dubbed the “Wisconsin Plan”—soon became a vital tool in fighting<|endoftext|>Books

4. #4 peak

   Breakfast replenishes the stores of nutrients and energy in your body. In the morning, after having not eaten for 8-12 hours, your glycogen stores are low. Eating breakfast boosts your energy levels so that your glycogen levels

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.