institution place nouns like school
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. 18 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them 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 name of a place where people work or study, such as 'school', 'university', 'campus' or 'district', especially after 'at the' or 'on the'. Turning it up made the model repeat phrases about a library and a track, and turning it down produced number-heavy 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.
-
The pumpkin patch will be watered and pampered until fall when each pumpkin will then be boiled, baked, and blended in the school kitchen and offered to students at lunchtime or during one of the school’s monthly
-
university's CMS, includes all web pages, documents, and other files made available via the university web site. The CMS makes it possible for almost anyone to easily manage web content
-
30, 2021 02:03 PM⏎Professor of Biology and Director of Sustainability Studies Rich Niesenbaum manages plants at the College's new pollinator habitat.⏎Summer has arrived in the Lehigh Valley, bringing
-
recording their heart rates and blood pressure during varying exercise intensities. The very nature of this activity meant that we had to leave the standard classroom and venture into the school gymnasium.⏎With this is mind I wanted to ensure that the data that
-
school district has a list of bus rules and regulations posted on the district website to assist parents and staff in making sure students are safe. OSD bus rules
-
fertilizer. This fertilizer is used to promote the lemon orchard located in the ACET campus. ACET discourages use of plastic; particularly single use plastics in campus. Paper wastes from departments, Library, Administrative offices, Hostels, are disposed
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 number of the person with the first one, one, and then the second person to the first, one, and then the third person to the first, and the process continues, and the sum of the first three numbers of the list
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
when you place an order online. Remember, if you feel uncomfortable providing any of this information online, please feel free to call one of our representatives at 1-800-WWW-DELL.⏎How Safe is Encryption Really?⏎Provid
-
95-9969. Questions or concerns relating to the accessibility of the FCC's online public file system should be directed to the FCC at 888-225-5322, at 888
-
hot or cold foods or beverages, you may have acid erosion. To schedule a dental checkup with Greg Ritchie at Ritchie Dental Group, call 830-693-8833 today. We are a full
-
. If you have any questions about your own oral health, or that of a loved one in your care, please contact Dr. Shane McDowell at 239-936-
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
- L5:2310 suppresses
- L8:84 suppresses
- L5:2634 excites
- L4:2557 weakly suppresses
- L15:2809 weakly suppresses
- L17:2786 weakly excites
- L10:552 weakly excites
- L14:1434 weakly suppresses
- L8:2585 weakly suppresses
- L6:313 weakly suppresses
- L1:215 weakly excites
- L9:763 weakly excites
- L9:1338 weakly excites
- L16:2287 weakly suppresses
- L10:2369 weakly suppresses
- L16:832 weakly suppresses
- L16:2627 weakly suppresses
- L11:1941 weakly suppresses
- L15:2376 weakly suppresses
- L16:1243 weakly suppresses
- L0:2167 weakly suppresses
- L17:978 weakly excites
- L6:649 weakly excites
- L9:2502 weakly suppresses
Feeds later layers
- L19:817 weakly suppresses
- L24:2083 weakly suppresses
- L20:1014 weakly excites
- L23:2534 weakly excites articles after building and room nouns
- L23:1624 weakly suppresses
- L21:2383 weakly suppresses
- L27:1613 weakly suppresses
- L21:2254 weakly excites
- L21:2315 weakly excites
- L24:2823 weakly suppresses
- L22:706 weakly suppresses
- L20:1163 weakly excites
- L26:137 weakly excites
- L20:1063 weakly excites
- L20:2168 weakly excites
- L22:2235 weakly suppresses
- L26:202 weakly excites
- L20:1474 weakly excites
- L20:2686 weakly excites
- L20:2089 weakly excites
- L27:1834 weakly suppresses
- L25:1687 weakly suppresses
- L19:1970 weakly excites
- L22:1272 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
- 墙
- sembl
- *</
- lickr
- Authority
- ски
- Screens
- ffic
- urse
- ximity
Suppresses
- 中共
- ajs
- 国人
- 盎
- pose
- empt
- UNCT
- cpp
- ivant
- bist