opening parenthesis after copyright or year
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 open bracket that comes right before a year in citations and copyright lines, like 'Copyright (c) 2009'. Turning it up made the model's text drift into loose, unrelated talk; the steering is not clearly connected to this pattern.
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.
-
water, were the major factors shaping phytoplankton communities.⏎- Abstract views: 1777⏎- PDF: 842⏎Copyright (c) 2009 Silvia TAVERNINI, Daniele N
-
of creating a “thermostatic mechanism” that would counter-balance the dominant presence of electronic media in the lives of children.⏎How to Cite⏎Copyright (c) 2017 Robert Albrecht⏎Copyrights and publishing
-
||Classical and Quantum Gravity|⏎|State||Published - Jul 7 2003|⏎ASJC Scopus subject areas⏎- Physics and Astronomy (miscellaneous)<|endoftext|>Most computer users are familiar with the concept of spyware —
-
icide and Social Change in China⏎- Cite this article as:⏎- Phillips, M.R., Liu, H. & Zhang, Y. Cult Med Psychiatry (1999) 23: 25. doi:1
-
viable. A symbol or instrument found in the Tuturangiana Andala ritual of green bamboo, betel nut, black goat, and red flag.⏎Copyright (c) 2022 Proceedings Of International Conference On Communication Science⏎This
-
. Ben Cohen. "The Racial Geography of the Federal Death Penalty" Washington Law Review (2010).<|endoftext|>Last updated 2/27/15
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 need for a more sustainable energy policy, like the green energy transition, and the push for climate action to reduce emissions from carbon taxes, and the Paris climate goals, like the 2030 targets to combat climate change), the
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
can contribute to the recovery of butterfly populations, although there is a paucity of information on the long-term effects of management. Using eight UK data sets (9–21 years), we investigate changes in restoration success for (1)
-
.⏎Climate change will disproportionately affect California’s most disadvantaged and isolated communities, according to a recent report from the Pacific Institute.⏎By looking at a broad array of factors – from social indicators such as income and birth rates, to environmental ones such as tree
-
ology for hints that the Red Planet may have ever had conditions hospitable to life.⏎The startlingly clear photo, one of several new views released by NASA today (Aug. 8), shows Curiosity's heat shield after it was shed about
-
in 1880. William Sutherland was one of several Scots who explored the country in the 19th century. He ended a wandering life (he had earlier fought under Garibaldi in Italy) when he settled at Mil
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
- L5:330 excites opening parenthesis after a noun
- L8:354 excites opening parenthesis before a citation year
- L7:582 suppresses
- L3:57 excites
- L6:2715 suppresses
- L5:1130 excites
- L6:1295 suppresses
- L7:2445 suppresses
- L5:1682 weakly suppresses
- L6:1412 weakly suppresses
- L4:64 weakly excites
- L3:928 weakly excites
- L2:1841 weakly suppresses
- L6:2247 weakly excites
- L8:2006 weakly excites
- L4:46 weakly suppresses
- L6:2453 weakly suppresses
- L8:1623 weakly suppresses
- L2:2440 weakly suppresses
- L1:1517 weakly suppresses
- L6:1567 weakly excites colon introducing a summary point
- L2:2134 weakly suppresses
- L5:602 weakly excites
- L3:487 weakly suppresses
Feeds later layers
- L13:608 suppresses opening parenthesis starting an aside
- L12:1248 suppresses
- L16:2801 suppresses
- L17:1657 suppresses
- L12:878 suppresses
- L15:261 suppresses
- L12:1837 suppresses
- L17:1908 suppresses
- L11:1072 suppresses
- L17:2364 suppresses
- L12:1784 suppresses
- L10:1459 suppresses
- L10:1654 suppresses opening parenthesis before an acronym
- L10:271 excites
- L18:2081 suppresses
- L16:906 suppresses
- L14:1181 suppresses
- L10:1756 suppresses
- L16:2296 suppresses
- L14:1209 excites
- L14:1470 excites
- L16:2177 suppresses
- L17:1430 excites
- L15:2787 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
- THON
- ymes
- Simulator
- 发生
- amination
- uration
- 又好
- esus
- 排行
- imedia
Suppresses
- >i
- beit
- assuming
- ieri
- itionally
- 斯坦
- bé
- efficient
- śmy
- >j