Neuron Explorer

Qwen3-0.6B

'and' or 'or' joining list items

A neuron in layer 14 of Qwen3-0.6B. Its strongest snippets are very consistent: they mostly peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on the word 'and' or 'or' when it joins two nouns or short items in a pair, like 'food and drink' or 'ceramics and composites'. When it was turned up, the model's text drifted into lists and repeated numbering.

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 activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

  1. #1 peak

    experimentally derived films. Good agreement between model and experiment was observed.⏎ASJC Scopus subject areas⏎- Electronic, Optical and Magnetic Materials⏎- Ceramics and Composites⏎- Condensed Matter Physics⏎- Materials Chemistry<|endoftext|>Humans are intimately

  2. #2 peak

    ⏎Stores other vitamins and minerals (fats, iron, copper, A, D, K, and

  3. #3 peak

    . In addition to vocabulary, there are also sentences and idiomatic expressions. New Amigos also includes geographical information and cards focused on culture, business, and food and drink.⏎New Amigos is a great game<|endoftext|>How did the kindly Christian saint

  4. #4 peak

    's Revised Unabridged Dictionary1. (n.) A thought or a consideration offered in support of a determination or an opinion; a just ground for a conclusion or an action; that which is offered or accepted as an explanation; the efficient cause

  5. #5 peak

    achieve an appropriate ‘fit’ with this environment. This includes:⏎Results that meet the expectations of its owners (shareholders, government, members etc. )⏎Products and services that meet its clients’ expectations at least as well as rivals’⏎Ability

  6. #6 peak

    direct link between eating nourishing foods and higher grades, alertness, and better memory and recall.⏎One simple reason for this is that ensuring you consume protein, carbs,

Steering

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.

turned down a lot
Prompt

Here is something I have been thinking about lately: what if we are trying to create a more sustainable future, and we want to build a better world, and we want to bring the attention to the importance of this, that we are to be our own leaders, but we want to have

Most negative on

Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.

  1. #1 peak

    une-system⏎- Kennedy, David O. “B Vitamins and the Brain: Mechanisms, Dose and Efficacy—A Review.” Nutrients, vol. 8, no. 2, Jan. 201

  2. #2 peak

    . (2021). Contrasting response of native and non‐native plants to disturbance and herbivory in mountain environments. Journal of Biogeography, 48(7), 1594-1605

  3. #3 peak

    to finance its assets. The WACC is the minimum acceptable return that a company must earn on an existing asset base to satisfy its creditors, owners, and other providers of capital, or they will invest elsewhere.⏎The following is the

  4. #4 peak

    Shale Revolution⏎- The Health Impacts of the Shale Revolution⏎- The Effects of the Shale Revolution on Local Governments⏎- The Shale Revolution and Water Quality⏎Common Resources — Feb 7, 2018⏎

Connections

Show on map

Which neurons in earlier layers feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.

Fed by earlier layers

  1. L6:798 excites
  2. L9:92 suppresses "and" or "or" joining two items
  3. L13:977 suppresses
  4. L13:131 excites 'and' joining parallel phrases
  5. L4:57 suppresses
  6. L11:406 suppresses
  7. L10:335 excites commas in catalog, date, and list metadata
  8. L6:427 suppresses
  9. L9:77 excites commas in citations and name lists
  10. L5:880 excites
  11. L4:3 excites commas after citation fields
  12. L7:2041 suppresses 'or' offering an alternative
  13. L7:836 excites
  14. L9:2010 suppresses
  15. L9:1942 excites
  16. L10:74 weakly excites
  17. L10:554 weakly excites
  18. L8:70 weakly excites 'and' between stacked superlatives
  19. L9:2244 weakly excites
  20. L12:1786 weakly excites
  21. L12:1235 weakly suppresses commas inside numbers and lists
  22. L4:2270 weakly excites
  23. L6:912 weakly suppresses
  24. L7:1223 weakly suppresses "and" joining two decades or times

Feeds later layers

  1. L21:735 excites 'or' in ranges and alternatives
  2. L15:1501 excites "and" joining two parts of a name or phrase
  3. L17:572 excites
  4. L15:1162 excites
  5. L17:2274 excites
  6. L15:2831 excites
  7. L17:1214 excites
  8. L15:734 excites
  9. L17:2386 excites
  10. L18:2136 excites
  11. L18:870 excites
  12. L16:1179 excites
  13. L15:2722 excites "and" or "or" in a pair or list
  14. L18:1270 excites
  15. L19:883 excites
  16. L16:340 excites
  17. L16:2580 excites 'and' linking food or health nouns
  18. L16:300 excites commas inside lists of nouns
  19. L16:1774 excites
  20. L18:1386 excites unclear: words before 'and' or 'or'
  21. L16:1369 suppresses
  22. L18:1625 excites
  23. L19:1823 excites 'and' inside organisation or paired names
  24. L16:1074 excites

Activation histogram

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.

162 values between -0.73 and -0.372304 values between -0.37 and -0.011264 values between -0.01 and 0.3538 values between 0.35 and 0.7119 values between 0.71 and 1.0610 values between 1.06 and 1.4214 values between 1.42 and 1.7812 values between 1.78 and 2.149 values between 2.14 and 2.505 values between 2.50 and 2.867 values between 2.86 and 3.214 values between 3.21 and 3.576 values between 3.57 and 3.938 values between 3.93 and 4.295 values between 4.29 and 4.656 values between 4.65 and 5.011 values between 5.01 and 5.374 values between 5.37 and 5.726 values between 5.72 and 6.086 values between 6.08 and 6.446 values between 6.44 and 6.803 values between 6.80 and 7.163 values between 7.16 and 7.521 values between 7.52 and 7.871 values between 7.87 and 8.231 values between 8.23 and 8.591 values between 8.59 and 8.951 values between 9.67 and 10.031 values between 13.25 and 13.61offnegativepositiveoften

Logit lens

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 layers instead.

Promotes

  1. /or
  2. phans
  3. equival
  4. isons
  5. consist
  6. acles
  7. combinations
  8. nbsp
  9. even
  10. дет

Suppresses

  1. “(
  2. Stories
  3. орт
  4. ariant
  5. amework
  6. ughter
  7. elic
  8. etration
  9. Award
  10. ascular