just before a pronoun for God

What this neuron does

It fires in religious writing about God, at points where the next word is likely to be 'He', 'His' or 'Him'. It seems to hold those pronouns back, and when it was turned down the model got stuck repeating 'He'.

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

   for Abraham, Isaac, and Jacob to keep, and yet they did the will of God because of their relationship with God. They knew His voice, and obeyed His voice as if what they heard was a law. God once told Jacob (

2. #2 peak

   40 day fast and temptations in the wilderness.⏎I. God speaks and His voice is very powerful⏎One of the wonderful things about our God is that He is a living relational God, who speaks to His creation.⏎Psalms

3. #3 peak

   of the names of God, but also applies equally to His servants. “The mu’min is the mirror of the Mu’min,” means that God unveils Himself in the mirror. Mevlana says when the sun shines in a mirror

4. #4 peak

   (1 Timothy 1:18). God continues to use children today in their schools, homes, communities, and nations to testify of His salvation and to accomplish His purposes. Many children in developing nations are at a loss as to who they

5. #5 peak

   why the Father sent His Son. God performed a miracle (the Incarnation) not as an end to itself, but as a means to allow us to be near Him in worship (through Christ’s redemptive sacrifice for our sins). Jesus

6. #6 peak

   an (perfection of worship), as detailed in a hadith: “Ihsan is to worship Allah as if you see Him; if you can’t see Him, surely He sees you.” Sufis regard Muhammad as al-Ins

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

   electricity for re-use by the drive train. Other vehicles capture the regeneration braking as a kinetic force to be released later to assist the powertrain or other under-the-bonnet components.⏎A good example of this new braking system is in the Toyota

2. #2 peak

   spliced⏎together different proteins can arise<|endoftext|>Ratio Forum is back on June 11 at Sofia Tech Park! Get ready for a full day of mind-blowing science facts and awesome innovations – all presented in a way accessible to all

3. #3 peak

   the sentence "necesito ayuda" "I need help." Or, a fluffly feline sits next to a glass of water, representing "tengo sed," or "I am thirsty."⏎The app also challenges its students to

4. #4 peak

   8 or more hours⏎Severe stomach or belly pain⏎Vomited three or more times in 1 day⏎Signs of dehydration include:⏎- Crying without tears⏎- Dry mouth⏎- Increased thirst⏎- Eyes that appear

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.