Neuron Explorer

Qwen3-0.6B

position right before a number

A neuron in layer 18 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron tends to fire on the spot just before a number, such as a blank where a figure comes next, or on a full stop before a name's initials. Turning it up produced rambling text that did not clearly relate to numbers.

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

    of approximately 1.7 million kilometers (1.1 million miles) from Prometheus and at a Sun-Prometheus-spacecraft, or phase, angle of 34 degrees. Image scale is 10 kilometers (6 miles) per

  2. #2 peak

    = vsin(theta) Immediately after take off : vx = 2cos25 = 1.8126 vy = 2sin25 = 0

  3. #3 peak

    ant.⏎Remove dried stems only after foliage dies back.⏎Rainy Side Notes⏎Lilium 'White Stargazer' is the most popular of the white upright lilies and

  4. #4 peak

    |⏎This species's range and population are probably declining as a result of habitat loss. It therefore qualifies as Endangered.⏎|Range Description:||Pyrrhura viridicata occurs only in the Sierra Nevada de Santa Marta

  5. #5 peak

    One Original Church⏎by the Most. Rev. Sarhad Jammo, Ph.⏎Ethnicity, Culture and Religion⏎Christian

  6. #6 peak

    oltage AC power supply⏎- Two capacitors, 0.1 µF each, non-polarized (Radio Shack catalog # 272-135)⏎- Two 27 kΩ resist

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: the [401 Pill" will be a way to be... 401 Pill is a 401 penny contribution, and the company's 2025-401PP is 521A-

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

    05-2009 / Preservative-Free Flu Shots⏎03-04-2009 / Flu Strains Developing Resistance to Tamiflu⏎03-03-2009 / The History Of

  2. #2 peak

    , which specifies where the runtime can locate an assembly (for more information, see Specifying an Assembly's Location).⏎<codeBase version="188.8.131.52" href="http://www.litware

  3. #3 peak

    5⏎Paperback ISBN: 9781503603844⏎Ebook ISBN: 9781503604513⏎Request Review/Desk/Examination Copy

  4. #4 peak

    He was recently given the honor of asking Edward Snowden the first question at his SXSW appearance where he talked about Internet privacy.<|endoftext|>Disclosure :: This post is sponsored by IDEA Public Schools. When we published our first post on IDEA Public Schools we received

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. L17:2257 suppresses
  2. L17:1336 suppresses
  3. L14:1796 suppresses
  4. L14:2157 suppresses
  5. L15:2509 suppresses
  6. L15:1873 weakly excites
  7. L12:398 weakly suppresses
  8. L15:777 weakly excites
  9. L15:949 weakly suppresses
  10. L9:1402 weakly suppresses
  11. L17:2635 weakly excites
  12. L16:1387 weakly suppresses
  13. L14:2037 weakly suppresses
  14. L10:2484 weakly suppresses
  15. L17:2621 weakly excites
  16. L2:103 weakly excites
  17. L10:2317 weakly excites
  18. L8:17 weakly suppresses
  19. L14:1436 weakly suppresses 'and' linking a role to a name
  20. L16:2217 weakly suppresses
  21. L17:2287 weakly excites
  22. L17:1860 weakly suppresses
  23. L13:1483 weakly excites
  24. L13:2256 weakly suppresses

Feeds later layers

  1. L19:2411 suppresses
  2. L19:929 suppresses
  3. L20:1174 suppresses
  4. L19:1607 suppresses
  5. L25:2383 weakly suppresses
  6. L19:873 weakly suppresses
  7. L24:710 weakly suppresses
  8. L19:2780 weakly excites
  9. L21:2847 weakly excites
  10. L20:38 weakly excites
  11. L20:191 weakly suppresses
  12. L19:1844 weakly suppresses
  13. L20:1361 weakly suppresses
  14. L25:1421 weakly suppresses
  15. L20:1232 weakly excites
  16. L27:2073 weakly suppresses
  17. L23:487 weakly suppresses
  18. L23:360 weakly suppresses
  19. L27:2694 weakly suppresses
  20. L20:1572 weakly suppresses
  21. L21:2948 weakly suppresses
  22. L19:586 weakly suppresses
  23. L25:531 weakly suppresses
  24. L21:557 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 -4.75 and -4.473 values between -3.91 and -3.632 values between -3.35 and -3.071 values between -3.07 and -2.793 values between -2.79 and -2.513 values between -2.51 and -2.237 values between -2.23 and -1.959 values between -1.95 and -1.676 values between -1.67 and -1.3917 values between -1.39 and -1.1127 values between -1.11 and -0.8361 values between -0.83 and -0.55286 values between -0.55 and -0.271474 values between -0.27 and 0.011329 values between 0.01 and 0.29301 values between 0.29 and 0.56117 values between 0.56 and 0.8475 values between 0.84 and 1.1242 values between 1.12 and 1.4030 values between 1.40 and 1.6825 values between 1.68 and 1.9624 values between 1.96 and 2.2415 values between 2.24 and 2.5214 values between 2.52 and 2.8010 values between 2.80 and 3.084 values between 3.08 and 3.363 values between 3.36 and 3.645 values between 3.64 and 3.924 values between 3.92 and 4.202 values between 4.20 and 4.483 values between 4.48 and 4.762 values between 4.76 and 5.041 values between 5.04 and 5.321 values between 5.60 and 5.881 values between 6.15 and 6.43offnegativepositiveoften

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. 徒
  2. 涿
  3. addObject
  4. trusting
  5. 貌似
  6. ello
  7. smouth
  8. 嘴
  9. stressing
  10. persist

Suppresses

  1. .spi
  2. ",[
  3. ponse
  4. theano
  5. łe
  6. ethoven
  7. áo
  8. atonin
  9. حال
  10. bote