Neuron Explorer

Qwen3-0.6B

opening parenthesis after citation author

A neuron in layer 8 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 reacts to the open bracket that comes right after an author's name in a reference list, as in "Cuenca, C. (2019)." Turning it up made the output drift into odd, unrelated text, so the steering shows little about 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

    larval mantis shrimp behavior remains elusive, their ultraviolet vision highlights the extraordinary visual adaptations of miniature marine creatures.⏎CONTENT PROVIDED BY⏎Congdon, O. (n.d.). How is a mantis shrimp’s vision like<|endoftext|>Part 2

  2. #2 peak

    Vol 3118, Springer-Verlag, Berlin⏎Heidelberg New York⏎D., Lenoir, R., and McKenzie, N., (200<|endoftext|>1. About lithium⏎Lithium is a type of

  3. #3 peak

    photography studio that Dalmau opened in 1854.⏎Download data is not yet available.⏎How to Cite⏎Cuenca, C. (2019). The The “Animated Polystereoscope

  4. #4 peak

    ., Lo, C.-M., Chan, K. S.-C., Chan, D. W.-O., & Chung, K. K.-H. (2014, July). Effectiveness of a tiered intervention model in

  5. #5 peak

    .1080/13527266.2011.620764⏎BELK, R. (201

  6. #6 peak

    of the approach taken and the justifications for⏎this approach, the following documents are recommended:⏎D., Lenoir, R., and McKenzie, N., (2004) Accessibility from scratch :⏎how an open focus contributes to inclusive

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 relationship between a person and their environment is a social relationship, which is important to understand. This is a question that could help you to think about the different ways in which people can be connected to each other, and how their connections are

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

    cyon lotor (Carnivora: Procyonidae): evidence of infection control?⏎Parasites & Vectors volume 7, Article number: 405 (2014)⏎A large variety

  2. #2 peak

    BMC Musculoskeletal Disorders volume 4, Article number: 4 (2003)⏎The purpose of this study is to

  3. #3 peak

    .⏎"Purchasing Power Parity Exchange Rates for the Global Poor,"⏎American Economic Journal: Applied Economics,⏎American Economic Association, vol. 3(2), pages 137-66, April.⏎- Angus Deaton & Olivier

  4. #4 peak

    was declared a conservation area.⏎Neil burn Drive was created.⏎Canmore place was erected.<|endoftext|>ERIC Number: ED379961⏎Record Type: RIE⏎Publication Date: 19

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. L6:1470 weakly suppresses
  2. L7:2165 weakly suppresses
  3. L7:22 weakly suppresses
  4. L4:2915 weakly suppresses
  5. L6:252 weakly suppresses
  6. L6:2305 weakly suppresses
  7. L3:1824 weakly suppresses
  8. L7:2426 weakly excites
  9. L2:617 weakly suppresses 'result' and 'results from' in science prose
  10. L2:2950 weakly suppresses
  11. L7:1037 weakly excites
  12. L6:1783 weakly suppresses
  13. L7:2807 weakly excites
  14. L0:2721 weakly excites
  15. L7:402 weakly suppresses
  16. L7:1263 weakly excites
  17. L0:1895 weakly excites
  18. L2:2523 weakly excites
  19. L6:2672 weakly excites
  20. L7:1133 weakly suppresses
  21. L5:1452 weakly excites "Al-Aqsa" and "Lunar Module" name parts
  22. L0:2197 weakly suppresses
  23. L3:482 weakly suppresses
  24. L7:1059 weakly suppresses

Feeds later layers

  1. L9:71 weakly excites
  2. L21:384 weakly suppresses
  3. L11:600 weakly excites
  4. L15:1488 weakly excites
  5. L26:799 weakly suppresses
  6. L9:33 weakly suppresses
  7. L27:177 weakly suppresses
  8. L16:2922 weakly suppresses
  9. L21:135 weakly suppresses
  10. L12:1457 weakly suppresses
  11. L25:2913 weakly excites
  12. L15:1169 weakly suppresses
  13. L15:2112 weakly suppresses
  14. L14:1904 weakly suppresses
  15. L19:1399 weakly suppresses
  16. L14:3017 weakly excites
  17. L9:1286 weakly excites
  18. L25:212 weakly excites
  19. L27:355 weakly excites
  20. L10:2013 weakly excites
  21. L14:2376 weakly suppresses
  22. L26:1794 weakly suppresses
  23. L16:870 weakly excites
  24. L16:2880 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.

3 values between -0.66 and -0.623 values between -0.62 and -0.592 values between -0.59 and -0.5510 values between -0.55 and -0.528 values between -0.52 and -0.4811 values between -0.48 and -0.4519 values between -0.45 and -0.4135 values between -0.41 and -0.3839 values between -0.38 and -0.3467 values between -0.34 and -0.3090 values between -0.30 and -0.27126 values between -0.27 and -0.23155 values between -0.23 and -0.20197 values between -0.20 and -0.16267 values between -0.16 and -0.13300 values between -0.13 and -0.09307 values between -0.09 and -0.06346 values between -0.06 and -0.02395 values between -0.02 and 0.01339 values between 0.01 and 0.05263 values between 0.05 and 0.08240 values between 0.08 and 0.12205 values between 0.12 and 0.15135 values between 0.15 and 0.19115 values between 0.19 and 0.2387 values between 0.23 and 0.2655 values between 0.26 and 0.3028 values between 0.30 and 0.3325 values between 0.33 and 0.377 values between 0.37 and 0.4012 values between 0.40 and 0.447 values between 0.44 and 0.473 values between 0.47 and 0.512 values between 0.51 and 0.541 values between 0.58 and 0.611 values between 0.61 and 0.651 values between 0.65 and 0.681 values between 0.68 and 0.721 values between 0.72 and 0.75offnegativepositiveoften

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. udson
  2. ypy
  3. 咽喉
  4. 一起装修
  5. opleft
  6. ’on
  7. OnChange
  8. :///
  9. abar
  10. ::__

Suppresses

  1. contrasts
  2. 瞥
  3. iolet
  4. ucci
  5. veled
  6. ười
  7. olds
  8. лей
  9. 恭喜
  10. qed