Neuron Explorer

Qwen3-0.6B

period or digit inside section numbers

A neuron in layer 12 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 fires on the dot, digit or dash inside numbered references like section 8.C.8.1, figure 5-18 or version numbers. Turning it up made the model write lists and numbered headings, which fits the theme.

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

    IFICATION; LEGAL CLUSTERS; GEOGRAPHIC CLUSTERS; TRADE CLUSTERS; ENVIRONMENT CLUSTERS; OTHER FACTORS. I. Identification. 1. The Issue The …⏎BONN, Germany (

  2. #2 peak

    4.3 feet of gulf water by the end of the century.⏎That rate could swamp

  3. #3 peak

    re-exported goods might require some additional information.⏎Having understood so much about reexport trade and what are re-exportation, it is<|endoftext|>10.2.2 Restarting TCP/IP⏎TCP/IP is a peer-to-peer, connection-oriented

  4. #4 peak

    - 3 Pulse transformer⏎- 4 RF transformer⏎- 5 Audio transformer⏎- 6 Other types⏎- 7 See also⏎- 8 References⏎This is the most common type of transformer, widely used in electric power

  5. #5 peak

    V. That’s 10% of the intended voltage!⏎So what are the traditional

  6. #6 peak

    that all highly toxic or offensive vapors are scrubbed or adsorbed before the exit gases are released into the hood exhaust system (see section 8.C.8.1 on fume hood scrubbers).⏎It is necessary to check if

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 can make a space that is a square, or a cube or a tetrahedron or a sphere, but I don't need to have a lot of things to put in or take out from. (This is a statement that

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

    Intravitreal injections are the fastest-growing procedure in ophthalmology and in medicine in general (Video 20-10, Fig 20-17). The most common indications for injections include AMD, diabetic retinopathy

  2. #2 peak

    . Self-equalizing Anchor. A self-equalizing anchor will maintain an equal load on each individual point as the direction of pull changes (Figure 5-18). This is sometimes used in rappelling when the route must change left

  3. #3 peak

    not be considered medical advice...This answer should not be considered medical advice and should not take the place of a doctor’s visit. Please see the bottom of the page for more information or visit our Terms and Conditions.⏎Thanks for your feedback.⏎3

  4. #4 peak

    product-term line . A structure such as Figure 5.11 (c) is called programmable array logic , first introduced by Monolithic Memories

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. L11:2116 excites dots inside version and section numbers
  2. L5:1316 excites
  3. L9:799 excites
  4. L9:1979 weakly excites
  5. L11:965 weakly excites
  6. L11:1532 weakly suppresses
  7. L8:2035 weakly suppresses
  8. L9:1587 weakly excites
  9. L4:1285 weakly excites
  10. L10:1516 weakly excites
  11. L10:2284 weakly suppresses
  12. L9:875 weakly excites
  13. L10:1479 weakly excites
  14. L11:2505 weakly excites
  15. L3:972 weakly suppresses
  16. L6:595 weakly suppresses
  17. L3:2466 weakly excites
  18. L11:869 weakly excites 'body' in immune and pain explanations
  19. L11:1685 weakly excites
  20. L7:1707 weakly suppresses
  21. L5:1363 weakly excites
  22. L4:1721 weakly excites the gap before a decade or century
  23. L7:868 weakly excites
  24. L5:26 weakly excites

Feeds later layers

  1. L13:1797 excites
  2. L14:1530 excites
  3. L17:1135 excites
  4. L20:925 excites
  5. L17:2548 excites
  6. L17:756 weakly excites
  7. L18:1640 weakly excites
  8. L17:2775 weakly suppresses
  9. L21:2710 weakly suppresses
  10. L13:2123 weakly suppresses
  11. L14:2446 weakly excites
  12. L16:1652 weakly excites
  13. L17:2696 weakly suppresses
  14. L13:852 weakly suppresses
  15. L13:2498 weakly excites
  16. L20:2393 weakly excites
  17. L21:2299 weakly excites
  18. L14:2354 weakly excites
  19. L27:393 weakly excites
  20. L23:1295 weakly suppresses
  21. L21:2372 weakly excites
  22. L20:1798 weakly excites
  23. L16:2306 weakly suppresses
  24. L23:1769 weakly suppresses

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 -2.40 and -2.311 values between -1.12 and -1.031 values between -1.03 and -0.951 values between -0.69 and -0.612 values between -0.61 and -0.522 values between -0.52 and -0.4410 values between -0.44 and -0.3558 values between -0.35 and -0.27188 values between -0.27 and -0.18469 values between -0.18 and -0.10921 values between -0.10 and -0.011052 values between -0.01 and 0.07730 values between 0.07 and 0.16299 values between 0.16 and 0.24120 values between 0.24 and 0.3340 values between 0.33 and 0.4112 values between 0.41 and 0.501 values between 0.93 and 1.01offnegativepositiveoften

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. below
  2. separate
  3. below
  4. -step
  5. step
  6. Expl
  7. 步骤
  8. Overview
  9. nutshell
  10. overview

Suppresses

  1. anka
  2. ány
  3. 索取
  4. unker
  5. taskId
  6. imeType
  7. Neville
  8. TES
  9. ipur
  10. /books