bodily samples and medical tests
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. 7 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 lights up on words for body samples and medical checks, such as blood in a stool test, stool samples, and urine. It is hard to pin down beyond that, and the words it avoids are mostly about word counts. Turning it up produced rambling text that did not clearly relate to medicine.
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.
-
copy, and fecal occult blood test.⏎- Cervical cancer: Regular Pap tests are recommended for women starting
-
2 – 19 MPN). Thereafter, FC abundance increased slightly in fall and early winter (1.9 – 24 MPN). Fecal coliform abundance during the 2000 La Niña year was much
-
a well-kept garden will yield a return ten to fifteen times greater than would the same area and location if devoted to general farm crops. A bountiful supply of vegetables close at hand where they may be secured at a few moments' notice
-
gut microbes, find out who they are and—more importantly it turns out—how they are behaving in these patients at risk for sepsis.⏎They surveyed stool samples from 14 patients with long ICU stays, including six people who stayed
-
for vitamin B12 deficiency, its clinical manifestations and⏎the possible repercussions for Australia’s refugee population.⏎Full blood count and blood film, iron studies and haemoglobinopathy studies are⏎routinely performed for newly arrived refugees in Australia.
-
610 mg of Sodium in this food.⏎Vitamins - This food has both Vitamin A and Vitamin C. One serving contains 500 International Units of Vitamin A and 1.2 mg of Vitamin C.<|endoftext|>A twig
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 people who like the number 15, the number of people who like the number 25, and so on, and so on, and so on... and so on... and so on... But I don't
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
. You can export a word list with a word frequency count to import into other visualization tools. http://www.antlab.sci.w
-
want to investigate, choose your scale of measurement and enter data. Using descriptive statistics, you'll find the mean, median, mode and standard deviation, and create frequency distributions and related graphs.⏎Monday, Oct. 6, 6-8
-
frequencies next to other words, and analyze the overall theme of a text and its corpus. The following tools listed below will help you get started with building a word frequency list and using your text to visualize your data, for the most part, in
-
or two weeks, but the circumstances were stronger then us. That’s why our trip to Kazachstan was much<|endoftext|>Data mining, concordances, word frequencies, all these things can be done to analyze text and to display the results (
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
- L6:2667 excites
- L6:683 suppresses
- L6:920 suppresses
- L6:1212 suppresses
- L6:666 suppresses
- L6:922 suppresses
- L6:1035 suppresses
- L3:1808 excites
- L6:16 excites
- L6:1203 weakly suppresses
- L6:1320 weakly excites
- L6:18 weakly excites
- L6:1887 weakly excites
- L6:1976 weakly suppresses
- L4:2700 weakly suppresses
- L4:1356 weakly suppresses
- L6:12 weakly excites
- L0:1741 weakly suppresses
- L5:2797 weakly excites
- L5:912 weakly excites
- L6:2129 weakly suppresses
- L5:2003 weakly excites
- L6:2102 weakly excites
- L6:2229 weakly excites
Feeds later layers
- L10:2379 suppresses
- L10:1657 suppresses
- L10:2772 suppresses
- L10:1617 suppresses
- L19:1511 suppresses
- L21:1607 suppresses
- L8:3053 suppresses
- L8:125 excites
- L15:2943 excites
- L22:1451 suppresses
- L18:1739 suppresses
- L10:2161 suppresses
- L22:2489 suppresses
- L11:1076 weakly suppresses
- L19:3043 weakly excites
- L8:2699 weakly excites
- L11:159 weakly suppresses
- L8:2625 weakly suppresses
- L9:701 weakly suppresses
- L20:1966 weakly suppresses
- L12:1350 weakly excites
- L9:139 weakly suppresses
- L11:2985 weakly excites
- L10:127 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
- 代办
- 箍
- 盾
- y
- yclic
- travellers
- ตรวจ
- mith
- ём
- 움
Suppresses
- istributions
- distribution
- distributions
- -Disposition
- requently
- omial
- amespace
- изатор
- otypes
- 占有率