mid-phrase words in set expressions
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. 13 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 is unclear. It fires on words in the middle of set phrases like 'one reason or another', 'such and such' and 'Which of the following motivations best explains so-and-so', and it dips on specific place and time words like 'today' in news text. Turning it up or down 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.
-
this⏎“Respected chief guest, honorable guest of honor and all the dignitaries present here, good evening and I would like to welcome you all for ………….function.”⏎After a brief introduction, there is always the inauguration of the
-
there is no single correct answer but for which there are perhaps more justifiable answers and less justifiable answers. Something like “Which of the following motivations best explains so-and-so’s actions in such-and-such novel?” Instead of posing this question
-
anks in advance. kamesh. here is the code from Marcus Mock exam 2: following code what will be the output?⏎I added UBB code here - Mapraputa [This message has been edited by Mapraputa Is (
-
, we will delve into one of Mary Oliver’s poems, analyzing its themes, literary devices, and the impact it has on readers.⏎The Poem: [Insert Poem Title]⏎[Insert the full text of the chosen Mary Oliver poem here
-
Every administrator and politician speaks of “party politics” in villages, and even villagers are often heard saying, “There is too much ‘party’ in such and such a village”. The coming of elections gives fresh opportunities for the crystallization of parties
-
?⏎Or imagine a student who comes home from school announcing that "she had a great day because she got an A, did better than her best friend, or ... won the spelling bee." These accomplishments reflect a very different set of goals than those
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: :?? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
of the medical staff at Phoenixville Hospital. His office is located at 824 Main St., Suite 206.<|endoftext|>MEDIA RELEASE - "CLIMATE IMP
-
of the West<|endoftext|>Arsenic and Old Space: NASA to Announce New Form of Life on Earth?⏎Gizmodo is reporting that NASA's big announcement later today will be that a new form of life has been found in the harsh environment of
-
that binge drinking has reached epidemic proportions in China and argues that efforts to tackle the problem must address the country's unique drinking culture.⏎In this study, published online today in the journal Addiction, binge drinking was defined as consuming 50g or
-
passed from the hands of civilization to civilization, from generation to generation that still appears as vibrant today as the day it was struck. - (LC.025)<|endoftext|>The fur trade caused Native people to become partially dependent on European trade goods
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
- L9:218 suppresses
- L12:2935 excites
- L11:1970 excites
- L12:2423 excites
- L10:1342 suppresses
- L9:570 excites
- L11:1972 suppresses
- L12:1788 excites
- L12:2684 suppresses
- L12:146 suppresses
- L11:974 suppresses
- L10:80 excites
- L12:461 suppresses
- L8:1506 excites
- L12:40 excites
- L10:158 excites
- L11:2095 suppresses
- L9:558 suppresses
- L12:26 suppresses
- L11:1711 weakly excites
- L9:27 weakly suppresses
- L12:955 weakly excites
- L9:849 weakly suppresses
- L9:1781 weakly suppresses
Feeds later layers
- L15:2255 strongly suppresses
- L14:1490 strongly suppresses
- L15:2165 suppresses
- L15:2525 suppresses
- L19:2756 suppresses
- L17:2138 suppresses
- L14:2251 suppresses
- L14:2067 suppresses
- L16:721 suppresses
- L14:2136 excites
- L15:1641 excites
- L14:1844 suppresses
- L17:2460 suppresses
- L17:2936 suppresses
- L19:2157 suppresses
- L14:797 excites
- L17:2036 excites
- L17:2024 suppresses
- L14:1604 excites
- L16:730 excites
- L15:1828 suppresses
- L15:1088 suppresses
- L15:1950 excites
- L14:2144 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
- ECH
- iction
- mvc
- enco
- icted
- itive
- 贫
- |R
- ics
- pel
Suppresses
- yourselves
- "><?=$
- љ
- ...">⏎
- ście
- Whatever
- 讓他們
- ...\
- )...
- ...');⏎