email address endings in examples
A neuronOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in layerOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. 9 of Qwen3-0.6B. Its strongest snippets are very consistent: they mostly peak on the same token. It is featuredA neuron picked for the full treatment: steering outputs and a written explanation. About 3,000 of the 86,000 neurons are featured., so it has steering outputs.
What this neuron does
This neuron lights up on the end of fake email addresses and web addresses, like the part right after the at sign in a placeholder email. When it was turned up, the model's writing mostly became repetitive or drifted into odd loops rather than showing a clear 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 activationThe number a neuron produces for a token. Large positive means it fired hard; negative means it pushed the other way. was highest. Amber marks positive activation; the outlined tokenA piece of text the model reads at once: often a word, sometimes part of a word or a punctuation mark. is the peak. Hover a token for its strength.
-
how your school can get involved with Bush Builders contact firstname.lastname@example.org⏎Brothers Island tuatara⏎For just over six years, Wellington
-
firstname.lastname@example.org.<|endoftext|>The Musical Representation⏎How human musical experience emerges from the audition of
-
the publication for benchmarking.⏎To explore opportunities for collaboration with subject-matter experts on your assessment and evaluation efforts, contact PRA Well-Being at firstname.lastname@example.org.⏎This resource was first shared in 2019.<|endoftext|>Social
-
the Department of Construction, Chaoyang University of Technology, Taichung 41349, Taiwan (e-mail: firstname.lastname@example.org).⏎Cite:Ming-Ju Lee, Ming-Gin Lee, Y
-
wide variety of languages, catering for beginning students to background speakers.⏎Our Faculty Members⏎Mr John Norberry: email@example.com⏎Mrs Laura Wade: firstname.lastname@example.org⏎Quite often as Language teachers, we are asked why students should have to
-
8|⏎Article from a library⏎|Web site with no author|⏎The examples above are reproduced from https://owl.english.purdue.edu/owl/resource/747/22/⏎
SteeringAdding a push along a neuron while the model writes, to see how its output changes. Turning it up or down.
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.
Here is something I have been thinking about lately: The original problem: Let's say I am. I. J. K. Question asked by I. J. K. **Question** **Answer** **Answer** **Question** **Answer** **Answer** **Question
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
op icthealth.nl. Link: https://ictandhealth.com/ucla-looks-into-rewiring-brain-in-order-to-help-with-depressions/news⏎UCLA looks into rewiring brain in order to help with depressions⏎
-
provider:|| Postal: Katajanokanlaituri 6B, 00160 Helsinki|⏎Web page: http://www.wider.unu.edu/⏎More information through EDIRC⏎Please report citation or reference errors
-
2018;15(2):57-61 DOI:http://dx.doi.org/10.4314/aas.v15i2.5⏎© 2018 Author. This
-
.org for more information. Updated June 25, 2015. (taken from http://genealogy.about.com/od/surname_meaning/p/boyle.htm ) BOYLE Last Name Meaning & Origin: A variant of
Connections
Show on mapWhich neuronsOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in earlier layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.
Fed by earlier layers
- L5:79 suppresses
- L3:254 suppresses
- L1:70 excites
- L2:25 suppresses
- L2:464 suppresses
- L3:21 excites
- L0:36 excites
- L3:171 excites
- L3:871 suppresses
- L4:58 excites
- L0:45 excites
- L0:29 suppresses
- L2:9 suppresses sentence-final periods vs dots in code
- L2:1775 excites
- L8:143 suppresses
- L1:1079 suppresses
- L5:1304 weakly excites
- L8:687 weakly excites
- L2:1170 weakly suppresses
- L3:195 weakly suppresses
- L7:5 weakly excites
- L7:1302 weakly suppresses
- L7:1913 weakly excites
- L5:227 weakly suppresses
Feeds later layers
- L15:722 strongly suppresses
- L10:134 strongly excites
- L10:342 suppresses
- L14:817 excites
- L15:1057 excites
- L12:1740 suppresses
- L15:2259 suppresses
- L10:856 suppresses
- L12:142 excites
- L10:984 excites
- L10:909 suppresses
- L15:2425 excites
- L15:241 suppresses
- L10:250 weakly suppresses
- L17:1775 weakly suppresses
- L17:603 weakly excites inside organisation and product names
- L14:345 weakly excites
- L15:625 weakly excites
- L10:2083 weakly suppresses
- L14:254 weakly suppresses
- L10:1642 weakly suppresses
- L16:2296 weakly suppresses
- L20:494 weakly excites
- L10:2225 weakly suppresses
Activation histogramA chart of how often a neuron takes each activation value.
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 lensReading which output tokens a neuron pushes toward or away from, straight off its output weights, before the later layers have their say.
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 layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. instead.
Promotes
- 荆
- ipping
- Truman
- tom
- %"><
- ']}'
- ores
- UIT
- cent
- raphics
Suppresses
- ereal
- aved
- those
- iese
- uelve
- 那些
- 犇
- alyze
- épo
- وقع