opening paren before an acronym
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 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 opening bracket after a full name, just before its short form, like the "(" in "Graduate Record Examination (GRE)". When it was turned up, the written text did not clearly change in any way that matches 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.
-
- In search of their lost ones.<|endoftext|>The Verbal Reasoning section of the Graduate Record Examination (GRE) challenges a student’s reading comprehension, vocabulary, and sentence completion skills.
-
numbers. The app will output the dataset's stem-and-leaf plot once Submit is clicked.<|endoftext|>Americans with Disabilities Act (ADA)⏎The Americans with Disabilities Act (ADA) provides civil rights protections to individuals with disabilities similar to those provided to individuals
-
an intact fibula<|endoftext|>Occupational Safety and Health Administration⏎OCCUPATIONAL SAFETY AND HEALTH ADMINISTRATION⏎The mission of the Occupational Safety and Health Administration (OSHA) is to assure that every U.S. worker goes home whole and
-
Social Cognition in Non-human Primates Using a Network of Computerized Automated Learning Device (ALDM) Test Systems⏎Institutions: Aix-Marseille University.⏎Fag
-
debt to women in dictatorships.”⏎Now, the CIDH has brought the ESMA exhibit to the Organization of American States (OAS) Headquarters in Washington DC in an attempt to start a conversation about crimes
-
- Reduce cholesterol and triglyceride levels⏎- Lower blood pressure⏎- Improve memory and cognition⏎What about cinnamon and HIV⏎Acquired immunodeficiency syndrome (AIDS) is caused by HIV and HIV-1 protease plays an essential
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 same. the same. the same. the same. the same. the same. the same. the same. the same. the same. the same. the same. I have been thinking about, the same. the same.
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
://ecf.hq.eso.org/newsletter/stecf-nl-22/freud
-
and a copy of the report go to the NOAA news website www.noaanews.noaa.gov⏎– Erich Luening⏎Print this page<|endoftext|>Today is Sunday,
-
and numbers. It worked by filling slots in frames maintaining an agenda of resource-limited prioritised tasks. AM's successor was Eurisko. http://homepages.enterprise.net/hibou/aicourse/lenat.txt.⏎Last updated:
-
Animal Adaptations". Please register for the second session through Eventbrite as well, if you are interested in attending both: https://scisatadapt2.eventbrite.ca⏎Please contact us if you have any questions & we hope to
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
- L9:792 strongly excites
- L2:504 strongly suppresses
- L0:670 suppresses
- L4:300 excites
- L9:2661 excites
- L8:143 suppresses
- L10:321 suppresses
- L12:1058 suppresses
- L5:1304 excites
- L2:1473 excites
- L7:366 excites
- L6:972 excites "well" after "how"
- L10:1979 excites
- L8:386 suppresses
- L0:860 excites chinese 的 particle before a noun
- L5:1851 weakly excites
- L0:1523 weakly excites
- L3:194 weakly suppresses
- L0:339 weakly suppresses
- L3:2168 weakly excites
- L7:1560 weakly suppresses
- L9:1009 weakly suppresses
- L1:1215 weakly excites
- L9:189 weakly excites
Feeds later layers
- L15:526 excites
- L17:1999 excites
- L15:779 suppresses
- L18:1129 suppresses
- L14:935 excites
- L16:2623 excites
- L15:309 suppresses
- L15:2616 suppresses
- L18:854 excites
- L20:2278 excites
- L22:1425 excites
- L15:2642 excites
- L19:2911 excites
- L25:2331 excites
- L24:140 excites
- L16:1433 weakly suppresses
- L14:2485 weakly excites
- L18:1575 weakly excites
- L21:2163 weakly excites
- L16:1594 weakly suppresses
- L16:1217 weakly suppresses
- L17:1996 weakly suppresses
- L17:2202 weakly excites
- L19:274 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
- 的过程中
- 的真实
- 地区的
- 产品的
- 的时间里
- 中的
- 里的
- 的真实性
- .db
- 的地
Suppresses
- changer
- raction
- ermal
- urret
- ategorical
- asio
- arent
- uchs
- udent
- واء