Neuron Explorer

Qwen3-0.6B

'of' in journal names

A neuron in layer 18 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 "of" inside the names of academic journals, such as "Journal of", "International Journal of" and "American Journal of". It is pushed the other way by minus and plus signs in maths problems. Turning it up did not produce a clear theme in the written 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 activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

  1. #1 peak

    E., Brown, S., & Ekman, S.L. (2002). Caregiver singing and background music in dementia care. Western Journal of Nursing Research, 24(2), 195-21

  2. #2 peak

    Dispersal of Aedes Aegypti in an Urban Area After Blood Feeding as Demonstrated by Rubidium-Marked Eggs," American Journal of Tropical Medicine Hygiene⏎52 (2):177-17

  3. #3 peak

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

  4. #4 peak

    . (2012). Social connections and the persuasiveness of viral campaigns in social network sites: Persuasive intent as the underlying mechanism. Journal of Marketing, 18(1), 39-53. http

  5. #5 peak

    sugars sometimes.⏎3. Green tea.⏎The World Journal of Gastroenterology did research and proved that green tea destroys some bad fats,

  6. #6 peak

    10.001⏎BERISHA-NAMANI, M. (2013). Information technology, internet, and marketing. International Journal of<|endoftext|>When Jesus lived in His fleshly body, He was alive to the power

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: I had a new customer who was a full-time employee of a client who was a major client of a third-time customer who was a client of a customer (a new customer herself) who was a complete customer of a customer (a new

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

    1 where n is the number of bits that represent the pixel. Thus, an 8-bit pixel would have a maximum value of (2^8) - 1 = 255. Let the maximum value for pixels in image

  2. #2 peak

    are perpendicular.⏎Let's see how to identify parallel and perpendicular lines.⏎Let's take some examples to understand this.⏎Considering two equations:⏎y = 2x +3 and y = 2x+5⏎On comparing y = 2

  3. #3 peak

    that,⏎Lines with the same slope are parallel.⏎Hence, the given lines are parallel.⏎Let's consider another set of equations.⏎y = 2x +3 and 4y = –2x+10⏎On comparing y

  4. #4 peak

    If we want to iterate an array in reverse order, we need to create a sequence for the for loop using range with:⏎- start as length of the array - 1⏎- stop as -<|endoftext|>12 April 1953

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. L8:1585 suppresses
  2. L14:446 weakly excites
  3. L15:1751 weakly excites
  4. L9:7 weakly excites
  5. L13:132 weakly excites "of" after a noun
  6. L7:117 weakly suppresses
  7. L0:1620 weakly excites
  8. L2:2546 weakly suppresses
  9. L15:310 weakly suppresses
  10. L11:2882 weakly suppresses
  11. L14:3030 weakly excites
  12. L11:936 weakly suppresses
  13. L8:18 weakly suppresses
  14. L3:2928 weakly excites
  15. L14:832 weakly excites modifier words before their head word
  16. L12:470 weakly suppresses
  17. L8:1032 weakly excites
  18. L1:2346 weakly suppresses
  19. L10:33 weakly suppresses
  20. L10:2522 weakly suppresses
  21. L9:41 weakly excites
  22. L1:1591 weakly suppresses
  23. L14:1634 weakly excites
  24. L15:1869 weakly excites

Feeds later layers

  1. L19:1923 weakly suppresses
  2. L21:1697 weakly suppresses
  3. L26:566 weakly excites
  4. L23:2179 weakly suppresses
  5. L19:885 weakly excites
  6. L24:2894 weakly excites
  7. L20:718 weakly suppresses
  8. L23:1502 weakly suppresses
  9. L22:1942 weakly suppresses
  10. L20:2986 weakly suppresses
  11. L19:1808 weakly suppresses
  12. L21:1257 weakly suppresses
  13. L19:1872 weakly suppresses
  14. L19:190 weakly suppresses
  15. L20:1783 weakly suppresses
  16. L19:787 weakly excites
  17. L19:301 weakly excites
  18. L19:1028 weakly excites
  19. L19:1472 weakly suppresses
  20. L25:2821 weakly suppresses
  21. L20:1164 weakly suppresses
  22. L26:232 weakly suppresses
  23. L20:1365 weakly suppresses
  24. L21:440 weakly 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.

1 values between -7.39 and -6.933 values between -5.54 and -5.073 values between -5.07 and -4.615 values between -4.61 and -4.154 values between -4.15 and -3.686 values between -3.68 and -3.2217 values between -3.22 and -2.7521 values between -2.75 and -2.2929 values between -2.29 and -1.8358 values between -1.83 and -1.36114 values between -1.36 and -0.90230 values between -0.90 and -0.441741 values between -0.44 and 0.031500 values between 0.03 and 0.49122 values between 0.49 and 0.9621 values between 0.96 and 1.4211 values between 1.42 and 1.882 values between 1.88 and 2.355 values between 2.35 and 2.813 values between 2.81 and 3.272 values between 3.27 and 3.742 values between 3.74 and 4.201 values between 4.20 and 4.671 values between 5.13 and 5.591 values between 5.59 and 6.061 values between 6.06 and 6.521 values between 6.52 and 6.981 values between 7.45 and 7.911 values between 9.77 and 10.231 values between 10.69 and 11.16offnegativepositiveoften

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. istributed
  2. ums
  3. atter
  4. (form
  5. asz
  6. ubits
  7. STIT
  8. distributed
  9. usan
  10. enic

Suppresses

  1. own
  2. 完整性
  3. OWN
  4. 完好
  5. 专享
  6. 主动性
  7. 完整
  8. hecy
  9. 占有率
  10. 发现自己