'to' before an infinitive
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. 24 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 fires on the word 'to' when it comes before a verb, as in 'remains to be seen' or 'much work to be done'. Turning it up didn't produce a clear theme in the sample text.
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.
-
-ready anytime soon. But, as Rashid writes, it could be the start of something:⏎The results are still not perfect, and there is still much work to be done, but the technology is very promising, and we hope that in a
-
possible cause of paranoia.⏎A person suffering from paranoia will read a negative meaning from an innocent remark, will believe a certain person or people in general are “out to get him” and will stay preoccupied with unfounded doubts about the people around
-
Eaton||Maintainer|⏎Resources and communication⏎|Developer||VCS Repository Webview||http://www.octave.org/hg/octave|⏎|Required to build||GNU make|⏎|Required to build||a recent version of g++
-
into the abyss – all this has already been suggested by experts as an explanation.⏎Where does Sars-CoV-2 go in the brain?⏎It also remains to be clarified where the virus travels to the brain. John H. Morrison, professor
-
0 refugees to the shores of Cuba, Louisiana, Guatemala, Florida, and France between 1789 and 1809.2 Much remains to be known about the St. [End Page 37] Dominguan dias
-
- Where was Jesus born? (Matthew 2:1)⏎- Who was the king at that time? (Matthew 2:1)⏎- Who came to see Jesus? (Matthew 2:1)⏎- How did they know he
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: just was just re-introduce just was just, just was just, just was just... just was just... just was just... just was just... just was just... just was just... just was just... just was just... just was
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
the border checkpoint at San Ysidro, California. “En nombre del cielo, les pido posada,” sings one group. “Pues no puede andar mi esposa am
-
otherwise he lives a normal life.⏎in normal circumstances, under normal circumstances en temps normal⏎→ In normal circumstances I would have resigned immediately.⏎to continue as normal se poursuivre normalement⏎→ Life in Israel will continue as normal.⏎w
-
to overcome a myriad of hurdles. plants use it to sprout branches no one gets covered and each gets maximum amount of sunlight, and as<|endoftext|>Помогите пожаааалуйста, сам не справляюсьY.
-
“Circassians are like a prism that shows just how polarized Israeli society is,” said Chen Bram, an anthropologist<|endoftext|>All classifieds – Veux-Veux-Pas, free classified ads Website.One moment, please⏎When
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
- L21:853 strongly excites 'can' after 'and' or a noun subject
- L21:1521 strongly excites
- L20:2089 strongly suppresses
- L23:505 strongly suppresses
- L21:958 strongly excites
- L20:636 strongly suppresses
- L21:1657 suppresses
- L22:627 suppresses
- L23:257 suppresses
- L22:479 suppresses
- L18:1852 excites
- L20:2262 excites
- L19:545 excites
- L19:730 excites
- L23:373 suppresses 'to' after a noun or adjective
- L19:1769 excites
- L19:385 suppresses
- L20:2883 suppresses
- L23:30 suppresses
- L18:2303 suppresses
- L19:1678 excites
- L19:1208 suppresses
- L20:2527 excites
- L22:2258 suppresses
Feeds later layers
- L26:2432 strongly suppresses
- L27:2609 suppresses
- L27:964 suppresses
- L26:385 suppresses
- L27:412 suppresses
- L25:503 suppresses
- L25:1841 suppresses
- L27:427 weakly excites
- L27:565 weakly excites
- L25:529 weakly suppresses
- L27:461 weakly excites
- L25:810 weakly suppresses
- L25:347 weakly excites
- L25:566 weakly suppresses
- L26:1629 weakly suppresses
- L25:357 weakly suppresses
- L27:1059 weakly excites
- L25:108 weakly suppresses
- L27:1038 weakly suppresses
- L27:501 weakly suppresses
- L27:1899 weakly excites
- L27:2506 weakly excites
- L25:586 weakly excites
- L26:2741 weakly excites
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
- ਯ
- 醒了
- 酩
- akening
- により
- ład
- rounded
- większ
- ayo
- Touches
Suppresses
- mieć
- łożyć
- fazer
- ивать
- hacer
- сделать
- использовать
- być
- contar
- ывать