Neuron Explorer

Qwen3-0.6B

closing parenthesis ending an aside

A neuron in layer 13 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 closing bracket or dash that ends a short aside inside a sentence, such as '(usually)' or '(maximum)'. The steered text did not show a clear link to this pattern.

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

    LOOKUP (click here for more information).⏎What is a 2D lookup?⏎In many cases, you only look for a value in – let’s say – column A and return the

  2. #2 peak

    traffic with neither party incurring Internet Transit fees. We describe the process of peering, introduce (a lot of) terminology, and enumerate the top five motivations ISPs give for peering. The chapter

  3. #3 peak

    U.S. territory, but the Livermore atmospheric center is looking at how wind changes or rain could affect the spread of nuclear materials in Japan.⏎"(The center) has been activated to provide the U.S. government with a predictive modeling capability to

  4. #4 peak

    both of these Conventions are obliged to bring their national traffic laws in conformity with (one of) these Conventions. Both Conventions entail rules on, among others, traffic

  5. #5 peak

    the synergy of every team member.⏎n.大綱, 提要 [ summary or outline of a book, play, etc ]⏎note: 字根組合的概念是 "general view",可以用來

  6. #6 peak

    ), data 0 (the library globals), and rloc 0 (the data relocation table). By contrast, an application if a resource database with (usually) five resources: code 0, code 1, data 0, r

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 user's query is in the blank, which is in the end (or else, which one, actually?...[...], as the...[...], which is in the end, and which one is actually the

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

    the key challenges facing education around the globe and to promote visionary thinking about new education models.⏎The summit was opened by Her Highness Sheikha Mozah Bint Nasser al-Missned, chair of the foundation, who said, “

  2. #2 peak

    hours before commanding Admiral Chuichi Nagumo began launching strike aircraft, the minesweeper USS Condor spotted a midget submarine outside the harbor entrance and alerted destroyer USS Ward. Ward was initially unsuccessful in locating the target. Hours later, Ward fired

  3. #3 peak

    marker. Battleship U.S.S. Ohio (approx<|endoftext|>“Snow White,” the beloved, first, full-length animated feature film

  4. #4 peak

    all the steps, you will see one shortcut on the desktop. Cut it and paste it at the following path⏎Now open Run (ctrl+R) and type your command name(Here my command name is “DOSBox“) and

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. L12:2203 excites
  2. L10:1883 weakly excites closing parenthesis after a number or unit
  3. L8:1902 weakly excites
  4. L12:968 weakly suppresses 'of' after geographic nouns
  5. L12:2516 weakly suppresses
  6. L12:1398 weakly excites
  7. L8:2208 weakly excites
  8. L5:3016 weakly suppresses
  9. L12:72 weakly suppresses
  10. L5:539 weakly excites
  11. L10:3054 weakly suppresses
  12. L9:830 weakly suppresses
  13. L9:2352 weakly excites
  14. L4:1220 weakly suppresses
  15. L8:507 weakly suppresses
  16. L4:133 weakly excites
  17. L12:3036 weakly suppresses
  18. L0:2766 weakly suppresses
  19. L8:1847 weakly excites
  20. L10:1517 weakly excites
  21. L4:2132 weakly suppresses
  22. L10:1706 weakly suppresses
  23. L4:518 weakly suppresses
  24. L1:2530 weakly excites

Feeds later layers

  1. L16:2306 suppresses
  2. L18:1891 suppresses
  3. L15:1657 suppresses
  4. L14:1943 excites
  5. L19:891 weakly suppresses
  6. L24:473 weakly excites
  7. L15:1640 weakly suppresses
  8. L18:2889 weakly excites
  9. L18:1097 weakly suppresses
  10. L16:1856 weakly suppresses
  11. L16:2436 weakly suppresses
  12. L16:592 weakly suppresses
  13. L16:2024 weakly excites
  14. L20:564 weakly excites
  15. L14:44 weakly excites
  16. L24:168 weakly excites
  17. L21:2800 weakly excites
  18. L14:1410 weakly suppresses
  19. L15:485 weakly suppresses
  20. L27:1932 weakly excites
  21. L27:451 weakly excites
  22. L20:1178 weakly excites
  23. L24:1072 weakly suppresses
  24. L17:1186 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.

4 values between -0.61 and -0.5510 values between -0.55 and -0.4920 values between -0.49 and -0.4340 values between -0.43 and -0.3795 values between -0.37 and -0.31229 values between -0.31 and -0.25333 values between -0.25 and -0.18506 values between -0.18 and -0.12637 values between -0.12 and -0.06802 values between -0.06 and 0.00608 values between 0.00 and 0.06289 values between 0.06 and 0.13141 values between 0.13 and 0.1982 values between 0.19 and 0.2529 values between 0.25 and 0.3122 values between 0.31 and 0.3715 values between 0.37 and 0.434 values between 0.43 and 0.504 values between 0.50 and 0.568 values between 0.56 and 0.624 values between 0.62 and 0.683 values between 0.68 and 0.744 values between 0.74 and 0.803 values between 0.80 and 0.871 values between 0.87 and 0.933 values between 0.93 and 0.991 values between 0.99 and 1.052 values between 1.05 and 1.111 values between 1.11 and 1.172 values between 1.17 and 1.233 values between 1.36 and 1.421 values between 1.42 and 1.481 values between 1.60 and 1.671 values between 1.79 and 1.85offnegativepositiveoften

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. _:
  2. BASIS
  3. 第一天
  4. AML
  5. kea
  6. 的日
  7. 意義
  8. дать
  9. expectations
  10. 明天

Suppresses

  1. icut
  2. ifer
  3. crow
  4. ies
  5. ione
  6. CDATA
  7. utions
  8. berry
  9. istes
  10. ucha