Neuron Explorer

Qwen3-0.6B

colons in clock times

A neuron in layer 11 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

The neuron responds to the colon in times of day, like '02:16 pm' or '17:47 GMT', in dates and timestamps. On the other side it dislikes certain religious and political names such as 'God' or 'Allah'. Turning it up produced unrelated text, so steering shows no clear 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

    the north although it could be overhead or elsewhere. For your best chance of sighting an aurora, try to look during the hours around local midnight (22:00-02:

  2. #2 peak

    I guess, it's not the first time it's happened...<|endoftext|>Last updated on September 17th, 2022 at 02:16 pm⏎Synodontis catfish are one of the most popular species

  3. #3 peak

    subject is much smaller than a human. Essential oils are highly concentrated and should<|endoftext|>Monday, October 5, 1998 Published at 17:47 GMT

  4. #4 peak

    SeaWeb/FIS)⏎Seafood producers tackle ocean acidification at annual Summit⏎Monday, February 21, 2011, 23:50 (GMT + 9)⏎From 31 January to 2

  5. #5 peak

    11:07:11 EDT⏎- How Martian winds make rocks walkThu, 8 Jan 2009, 14:01:08 EST⏎- MIT moves toward greener chemistryFri,

  6. #6 peak

    'Lose my life, my leg or my hand?': Modern-day slavery in India⏎10 July 2014 Last updated at 18:48 BST⏎It is a form of modern day slavery otherwise known as forced

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 think the United States is the first, the second, the third, and so on, the highest, highest, highest, and so on, the highest, highest, highest, and so on, the highest, highest, highest

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

    of the nation" (White Paper on Tibet, Sep 1992). Chen Kuiyuan, a Chinese cadre appointed as leader of the Chinese Communist Party in Tibet in March 1992, has called on "in

  2. #2 peak

    107 of the Copyright Act 1976.⏎Fair son, the first thing I would teach thee is to set thine heart to love God; for unless he love God none can be saved.”⏎From Louis’ last instructions

  3. #3 peak

    is necessary for all new converts to recite this upon conversion to the new faith.⏎The first part of this declaration is that there is none worthy of worship except Allah. As we have previously mentioned, Allah is the personal name for the One God

  4. #4 peak

    the establishment of the People's Republic of Albania in January 1946, Albania became a rigid police state, dominated completely by the communist party and by Marxism/Leninism. Although Albania operated under the facade of constitutional rule, the

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. L4:42 suppresses
  2. L3:1262 weakly excites
  3. L10:2802 weakly suppresses
  4. L10:495 weakly excites
  5. L8:12 weakly excites
  6. L3:136 weakly excites
  7. L4:2224 weakly excites
  8. L9:824 weakly suppresses
  9. L9:1480 weakly suppresses
  10. L8:4 weakly excites
  11. L3:2269 weakly suppresses
  12. L8:2404 weakly suppresses
  13. L10:1921 weakly suppresses
  14. L4:1259 weakly suppresses
  15. L9:2160 weakly excites
  16. L2:283 weakly excites
  17. L8:3020 weakly excites
  18. L7:1826 weakly suppresses
  19. L8:2930 weakly excites
  20. L10:2014 weakly excites
  21. L6:1247 weakly excites
  22. L7:1962 weakly suppresses
  23. L5:758 weakly suppresses
  24. L9:2955 weakly suppresses

Feeds later layers

  1. L13:657 suppresses
  2. L14:1681 excites
  3. L16:2360 weakly suppresses
  4. L22:1726 weakly excites
  5. L14:1153 weakly suppresses
  6. L14:2632 weakly excites
  7. L17:1591 weakly suppresses
  8. L22:1518 weakly excites
  9. L15:2056 weakly suppresses
  10. L15:1556 weakly suppresses
  11. L21:1133 weakly suppresses
  12. L24:2309 weakly suppresses
  13. L23:1082 weakly excites
  14. L14:540 weakly suppresses
  15. L24:909 weakly suppresses
  16. L15:281 weakly suppresses
  17. L25:2731 weakly suppresses
  18. L14:2106 weakly excites
  19. L13:431 weakly suppresses
  20. L15:509 weakly suppresses
  21. L24:1346 weakly excites
  22. L13:1848 weakly excites
  23. L20:2036 weakly excites
  24. L23:356 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.

2 values between -1.98 and -1.891 values between -1.28 and -1.191 values between -0.67 and -0.593 values between -0.59 and -0.5011 values between -0.50 and -0.4138 values between -0.41 and -0.3392 values between -0.33 and -0.24334 values between -0.24 and -0.15667 values between -0.15 and -0.061316 values between -0.06 and 0.02837 values between 0.02 and 0.11393 values between 0.11 and 0.20149 values between 0.20 and 0.2854 values between 0.28 and 0.375 values between 0.37 and 0.463 values between 0.46 and 0.541 values between 0.63 and 0.721 values between 1.41 and 1.50offnegativepositiveoften

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. 展品
  2. extras
  3. út
  4. ojis
  5. habitat
  6. pró
  7. ерь
  8. erte
  9. ophile
  10. statt

Suppresses

  1. acknowledgement
  2. própria
  3. 葭
  4. yourself
  5. commercial
  6. onest
  7. 鸱
  8. otten
  9. 爱你
  10. orton