copyright and rights notices
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. 14 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. 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 responds to the end of copyright lines, such as 'All rights reserved' or '© 2018 Author'. These are the boilerplate lines that appear at the bottom of web pages and papers. Turning it up did not bring out a copyright theme in the 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 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.
-
trends in NCD prevalence, their risk factors, and treatments.⏎Copyright © 2020 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.<|endoftext|>We all grew up hearing the phrase, “Good night
-
1 DOI:http://dx.doi.org/10.4314/aas.v15i2.5⏎© 2018 Author. This work is licensed under the Creative Commons Attribution 4.0 International License
-
iquette and Responsibilities Brochure.<|endoftext|>Portuguese translation of 'discontent'⏎Copyright © 2014 by HarperCollins Publishers. All rights reserved.⏎Example Sentences Including 'discontent'⏎Marriage, discontent with home,
-
other nine were found to be “obstructive.”<|endoftext|>Rural earthen roads impact assessment in Phewa watershed, Western region, Nepal⏎© The Author(s). 2016⏎Received: 23 November 2
-
and Other Writings (1971).⏎The Columbia Electronic Encyclopedia, 6th ed. Copyright © 2012, Columbia University Press. All rights reserved.⏎See more Encyclopedia articles on
-
, China⏎2 Nanjing Xiaozhuang University, Nanjing, Jiangsu 211171, China⏎*Authors to whom correspondence should be addressed.⏎Received: 2020-8-4 / Accepted
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 effect of a small amount of information. The effect of a small amount of information. The effect of a small amount of information. The effect of a small amount of information. The effect of a small amount of information. The effect of
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
015 framework and its preparation if poverty is to be taken more seriously by donors.⏎|Date of creation:||2013|⏎|Date of revision:|⏎|Contact details of
-
last during His, not one another's pleasure.⏎This idea found its way into our Declaration of Independence in modified form:⏎We hold these truths to be self-evident, that all men are created equal, that they are endowed by their Creator
-
We hold these Truths to be self-evident, that all Men are created equal, that they are endowed by their Creator
-
Pakistan⏎||It has been suggested that this article be merged with Blasphemy law#Pakistan. (Discuss) Proposed since March 2
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
- L13:8 excites
- L10:169 weakly excites
- L4:181 weakly suppresses
- L13:2974 weakly excites
- L11:1475 weakly suppresses unclear: end of boilerplate text
- L10:2279 weakly excites
- L12:2177 weakly suppresses
- L12:2642 weakly suppresses
- L10:2818 weakly suppresses
- L12:405 weakly excites
- L13:1762 weakly suppresses
- L13:1277 weakly excites
- L2:62 weakly excites
- L10:2696 weakly excites
- L6:1672 weakly suppresses
- L9:2736 weakly suppresses
- L11:853 weakly excites
- L8:2628 weakly suppresses
- L1:533 weakly excites
- L4:564 weakly excites
- L3:795 weakly excites
- L2:2091 weakly excites
- L13:428 weakly suppresses
- L9:2863 weakly excites
Feeds later layers
- L16:1427 excites
- L18:1817 excites
- L20:309 excites
- L16:2247 weakly suppresses
- L19:2173 weakly suppresses
- L21:1985 weakly excites
- L15:2675 weakly excites
- L16:3064 weakly excites
- L15:2597 weakly suppresses end of a web document
- L15:1726 weakly suppresses
- L15:2861 weakly suppresses 'only' in 'not only ... but'
- L23:2395 weakly excites
- L25:381 weakly suppresses
- L20:1774 weakly excites
- L20:1539 weakly suppresses
- L21:2855 weakly excites
- L19:1758 weakly excites
- L17:1398 weakly suppresses
- L18:720 weakly excites
- L21:2170 weakly excites
- L18:412 weakly suppresses
- L18:1403 weakly suppresses
- L16:370 weakly suppresses
- L23:340 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
- WALL
- aña
- ron
- !⏎⏎⏎⏎⏎⏎
- tron
- ACHI
- !⏎⏎⏎
- achi
- Tri
- tical
Suppresses
- INESS
- stinence
- fict
- onders
- 肱
- stash
- estival
- 取证
- hypothetical
- evidence