'energy' in home energy topics
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. 8 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 'energy' in articles about home heating, insulation and energy bills. When turned up, the model's writing drifted toward houses, heat and energy costs, but this was only loosely connected.
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.
-
energy bills.⏎WHERE DOES MY HOME LOSE ENERGY?⏎Superior energy efficiency mostly depends on creating a better air seal. Wall insulation is a significant issue in home energy savings, but there are other factors that will lower home energy efficiency.⏎Doors
-
winter heat loss and summer heat gain. Tight construction, good insulation and high-performance windows are key. "Weather stripping and caulking in existing homes can reduce your energy bill 10 to 25 percent," says David Johnston, president of
-
of the air leakage in a home.⏎As much as 40% of energy loss is due to air infiltration. Therefore, the most effective way to improve energy efficiency and comfort in your home is to seal it with foam.⏎LET’S COMP
-
the effects of climate change. Reference Energy System (RES) is a method of analysis for certain systems in particular scenarios. In this study, we examine a house energy system modelling under different circumstances. Our aim is to increase energy saving and use more
-
foam installations are carried out by Lapolla / AirTight licensed installers.⏎*See manufacturer guidelines and requirements.⏎HOME ENERGY ASSESSMENT⏎A home energy assessment helps owners determine where their house is losing energy and money – and how such
-
Post your comments<|endoftext|>Green buildings are designed, constructed and operated for the purpose of providing the occupants with a space that is both healthy and productive, while also being energy efficient and sustainable. Unfortunately, current approaches to green building design and construction don’t
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: In a small town, a group of friends are in a cozy, rural town hall. The only living, they are still in the same town. The next day, another day, and another. **Plot Summary: **
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
explained that he suspected Samuel Robinson and the stranger as the real thieves. The authorities were not convinced. Pat Lyon was arrested and thrown in Walnut Street Prison.⏎Elsewhere in Philadelphia, Isaac Davis (the stranger Robinson had brought to the vault) began
-
he does it. There’s just one teensy, weensy problem.⏎Thomas never reached out and touched Jesus’ wounds. Not at all.⏎Meanwhile, in modern day Sunday school land, Thomas is the quick and easy way for teachers and
-
: Sam goes up as Hiroko zips down. And there, in the tree: one squirrel climbs up and another scurries down. Not far away, Daddy is quiet but Simon is LOUD, while a mother bird is silent as
-
to track the bear, but a snowstorm had obliterated the tracks. Search parties were sent out, but they found no trace of the dead cook. Meanwhile, Ryan made himself scarce while the men were out searching, and no one ever found
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
- L7:174 suppresses
- L6:2442 weakly suppresses
- L6:1832 weakly suppresses
- L6:1272 weakly excites
- L7:802 weakly excites
- L5:347 weakly suppresses
- L4:1589 weakly excites
- L5:1744 weakly suppresses
- L0:1603 weakly suppresses
- L6:2259 weakly excites
- L2:235 weakly suppresses
- L0:696 weakly excites
- L0:1832 weakly suppresses
- L4:2779 weakly suppresses
- L4:1579 weakly suppresses
- L5:3043 weakly suppresses
- L6:1537 weakly excites
- L3:2546 weakly excites
- L0:1293 weakly excites
- L0:2831 weakly excites
- L1:2922 weakly excites
- L3:1361 weakly excites
- L1:1889 weakly excites
- L6:881 weakly excites
Feeds later layers
- L11:917 excites
- L15:1552 weakly excites
- L17:2810 weakly excites
- L12:2872 weakly excites
- L20:2466 weakly excites
- L18:1816 weakly excites
- L14:2094 weakly suppresses
- L13:2979 weakly suppresses
- L10:1009 weakly excites
- L17:2737 weakly excites
- L15:2781 weakly excites
- L16:2899 weakly suppresses
- L9:1633 weakly excites
- L11:2213 weakly suppresses
- L14:2132 weakly excites
- L11:2950 weakly excites
- L20:1150 weakly excites
- L12:1738 weakly excites
- L20:2562 weakly excites
- L10:670 weakly excites
- L9:2290 weakly suppresses
- L25:2329 weakly excites
- L14:2655 weakly suppresses
- L12:1044 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
- ieten
- ensible
- 箅
- ��
- ','',
- ration
- /rfc
- enci
- [state
- ','');⏎
Suppresses
- serial
- orneys
- лав
- kode
- untos
- ultrasound
- standing
- INAL
- laughter
- hold