"heart" in medical and body words
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. 23 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 fires on the word "heart" when it appears in health contexts such as heartbeat, heart rate or heartburn. Turning it up or down produced rambling text with no clear link to hearts, so the steering shows little about it.
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 toxic components of lily of the valley affect your dog's heart. These components are called cardiac glycosides, and the symptoms they cause include irregular heartbeats, a slowed heart rate and seizures. Vomiting and diarrhea are
-
tremors, high blood pressure, irregular heartbeats, and extreme mood swings.⏎These are very scary side effects, particularly
-
body of a child can feel the effects for hours. In addition to difficulty sleeping, kids who drink too much caffeine may experience upset stomach, headaches, and increased heart rate and blood pressure. The products that contain caffeine often cause other problems, too
-
Females: white collar behind head. Males: White striations throughout body⏎||Atlantic Ocean and Gulf of Mexico coastal areas; Kansas; Oklahoma⏎||Heartwater, African Theileriases⏎|Lone Star Tick
-
who invented it back in the 19th Century, can be very effective in the right hands. It can determine the age, position and heart rate of the foetus, along with an indication of its overall health. But
-
say they are willing to stay longer and rate the products and stores as higher quality in a shopping district with the trees.⏎Trees in the landscape relax us, lower heart rates, and reduce stress.⏎Apart of what we do here W-Stats Tree
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 idea that a company’s financial and nonfinancial management (FFNM) training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training training
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
happening in our daily lives but the two practices combined may provide more objectivity from which to view these unpleasant things.⏎MBCT's main technique is based on the Mindfulness-Based Stress Reduction (MBSR) eight week program, developed by Jon
-
89%, a sensitivity 70% and 75% specificity. Conclusion: Use of CCHR would reduce unnecessary use of CT scans in minor head injury in this setup.⏎Keywords: Canadian, head
-
Canadian Head CT-scan rule (CCHR) in minor head injury is an evidence-based aid in decision making as regards to use of CT-scans to detect head injury requiring neuro-intervention. It therefore avoids wastage of resources. The objective
-
in the present, not just for oneself, but also for others. Ultimately, "Groundhog Day" gives us a lesson in mindfulness.⏎Metaphor for mindlessness?⏎Phil was trapped in Groundhog
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
- L18:2156 excites
- L1:424 weakly suppresses
- L22:2098 weakly suppresses
- L20:2805 weakly suppresses
- L5:2289 weakly suppresses
- L19:2064 weakly suppresses
- L22:151 weakly suppresses
- L19:647 weakly excites
- L22:911 weakly excites
- L11:930 weakly suppresses
- L22:3003 weakly suppresses
- L19:790 weakly excites
- L20:687 weakly suppresses
- L2:2927 weakly excites
- L4:2526 weakly excites
- L21:1178 weakly suppresses
- L3:958 weakly suppresses
- L2:2892 weakly excites
- L4:1347 weakly excites
- L7:952 weakly excites
- L16:2593 weakly excites
- L1:1712 weakly excites
- L22:2976 weakly suppresses
- L5:1874 weakly suppresses
Feeds later layers
- L27:909 excites
- L24:296 weakly suppresses
- L24:1101 weakly suppresses
- L24:1648 weakly suppresses
- L24:2540 weakly excites
- L27:2952 weakly excites
- L25:1706 weakly excites
- L24:2373 weakly suppresses
- L24:2039 weakly suppresses
- L24:1537 weakly excites
- L26:1090 weakly excites
- L26:2018 weakly suppresses
- L24:259 weakly suppresses
- L27:1128 weakly excites
- L24:2219 weakly excites
- L24:2560 weakly excites
- L25:3047 weakly excites
- L27:1934 weakly excites
- L24:508 weakly excites
- L26:444 weakly suppresses
- L27:203 weakly suppresses
- L24:2820 weakly excites
- L24:2128 weakly excites
- L26:1278 weakly excites
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
- ache
- ató
- ugar
- ador
- Beat
- /mac
- adora
- los
- achen
- üt
Suppresses
- hand
- ful
- brains
- FUL
- Ear
- fullscreen
- nees
- 大脑
- Hand
- 嘴巴