unclear: word fragments inside names
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. 4 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 fires weakly on pieces of longer words or names, such as the middle of 'Meiji' or 'Bemidji', and it is strongly negative on 'of' in journal citations. The pattern is not clear, and turning it up did not produce a clear theme.
Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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.
-
yo-e prints, which were mass-produced woodblock prints depicting everyday life and famous actors and courtesans. The Meiji period (1868 CE to 1912 CE) saw
-
.⏎UrbanDazzle has fantastic collections of water decanters, bottles and flasks. There are bottles by Tiger, Zojirushi that add an element of style, all the while keeping your<|endoftext|>Types
-
ESV⏎• Luke 9:58 NASB⏎• Luke 9:58 KJV⏎• Luke 9:58 Commentaries⏎• Luke 9:
-
DahM is also⏎defined as "dumb." In Leviticus 10:3 the suddenly bereaved Aaron⏎held his peace (KJV and the imitative old JPS), but VaYiDoaM⏎
-
of the giant planet Jupiter and its tiny lava-covered moon Io.⏎When the solar-powered spacecraft completed its 47th close pass (perijove) from Jupiter on December 14, it attempted to return its science
-
said Nancy Riestenberg, School Climate Specialist at the Department of Education.⏎Riestenberg joined members of the task force to speak with students from Bemidji High School and students from Schoolcraft to discuss their concerns about bullying and ways in
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: I need to be able to work with the other day, and I need to be able to work on the other day. I need to be able to work on the other day. So, I am writing this response.
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
Marini, L. (2021). Contrasting response of native and non‐native plants to disturbance and herbivory in mountain environments. Journal of Biogeography, 48(7), 1594-
-
. (2012). Social connections and the persuasiveness of viral campaigns in social network sites: Persuasive intent as the underlying mechanism. Journal of Marketing, 18(1), 39-53. http
-
., Spackman, E. 2008. Do crocodilians get the flu? Looking for influenza A in captive crocodilians. Journal of Experimental Zoology. 309A:1-10. Interpret
-
and make appropriate adjustments.:⏎- Andrew Shepherd, 2001. "Consolidating the lessons of 50 years of 'development'," Journal of International Development, John Wiley & Sons, Ltd., vol. 1
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
- L1:134 suppresses
- L0:2459 excites
- L0:244 weakly suppresses
- L1:2683 weakly excites
- L0:2693 weakly excites
- L2:1774 weakly excites
- L1:68 weakly excites
- L1:2032 weakly excites
- L1:92 weakly excites
- L2:1321 weakly suppresses
- L1:416 weakly suppresses
- L1:1931 weakly suppresses
- L0:1617 weakly suppresses
- L1:1973 weakly suppresses
- L3:1434 weakly suppresses
- L1:1464 weakly excites
- L3:2400 weakly suppresses
- L2:75 weakly suppresses
- L3:187 weakly excites
- L0:2589 weakly excites
- L3:50 weakly excites
- L2:1791 weakly excites
- L1:2819 weakly suppresses
- L2:1493 weakly excites
Feeds later layers
- L20:600 suppresses
- L19:1282 suppresses
- L21:561 suppresses
- L23:2855 suppresses
- L18:541 excites
- L5:2652 weakly suppresses
- L25:2231 weakly suppresses
- L25:2856 weakly suppresses
- L17:1313 weakly suppresses
- L5:1707 weakly suppresses
- L27:2703 weakly suppresses
- L26:1973 weakly suppresses
- L6:483 weakly suppresses
- L26:2903 weakly suppresses
- L8:1678 weakly suppresses
- L7:154 weakly excites
- L7:1924 weakly suppresses
- L25:2061 weakly suppresses
- L22:2628 weakly suppresses
- L19:2730 weakly suppresses
- L10:887 weakly suppresses the B in vitamin B names
- L27:2204 weakly suppresses
- L26:2556 weakly suppresses
- L27:2746 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
- over
- izers
- overs
- idea
- Over
- +
- ightly
- REST
- lords
- IP
Suppresses
- OURNAL
- กระบ
- ournals
- ####⏎
- 蜇
- synonym
- Braz
- <<⏎
- 같
- ę