Neuron Explorer

Qwen3-0.6B

comma after 'still' or 'in addition'

A neuron in layer 4 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 a comma that follows a short linking word at the start of a sentence, such as 'Still,', 'In addition,' or 'most importantly,'. Turning it up did not make the written text show any 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

    time frame to take complete responsibility for getting themselves ready, ask questions to help prompt them if you think they aren’t paying attention to something crucial, and most importantly, when things aren’t crucial (e.g. do they have the right uniform packed

  2. #2 peak

    treated several other parishioners in the same way and with the same result, and the treatment was widely adopted by 18th-century medics.⏎Still, Cartwright might have settled for rural obscurity in his rectory had it not

  3. #3 peak

    the spread of germs,⏎- Keeping common surfaces clean, and⏎- Staying home when you are sick.⏎Currently there are no advisories restricting

  4. #4 peak

    generalization.)⏎Your reader will be thankful if you start with “This question…” or “This wig…” or, most likely, “This intrusion…” If<|endoftext|>A newly built apartment building consists of several floors with

  5. #5 peak

    services and evaluate video quality in six categories of digital defects on a scale of one to five (one being unwatchable and five being spectacular).”⏎“In addition, Comcast is working to establish local golden-eye programs in at least two markets.”⏎“My

  6. #6 peak

    great psychological benefits, which may spur you on to other healthy behaviors such as making regular exercise part of your life, too.”⏎Can anyone try intermittent fasting?⏎Still, most experts agree that some people should steer clear of IF altogether. That includes kids

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: 12, 20, 14, 16, 24, 32, 40, 48, 50, 60, 64, 72,

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 this humongous ship, but many believe it had something to do with low grade rivets.⏎Wealthier passengers were offered luxurious dining tables with the best meals, but none of this

  2. #2 peak

    .Jack and the Beanstalk.The Seven Ravens.The Goose Girl.Brother and Sister A brother drinks from a stream and turns into a deer just when the

  3. #3 peak

    sprang. Amidst all the animosity in the play, it's easy to forget that the w

  4. #4 peak

    -set' of a clip with its own characteristics. Amongst other advantages, object-oriented editing can simplify the automation process compared to drawing track envelopes

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. L3:155 strongly excites
  2. L2:63 strongly suppresses
  3. L3:1081 strongly excites
  4. L1:454 suppresses
  5. L3:195 suppresses
  6. L3:1071 excites
  7. L3:45 suppresses
  8. L2:501 suppresses
  9. L3:226 excites
  10. L3:1093 suppresses
  11. L2:307 excites
  12. L3:472 suppresses
  13. L3:81 suppresses
  14. L3:194 excites
  15. L2:329 excites
  16. L1:1759 excites
  17. L3:246 excites
  18. L3:208 excites
  19. L2:25 suppresses
  20. L0:2921 suppresses
  21. L3:21 excites
  22. L1:1192 suppresses
  23. L3:15 excites
  24. L1:456 suppresses

Feeds later layers

  1. L14:817 excites
  2. L10:856 suppresses
  3. L5:219 excites
  4. L6:173 excites
  5. L5:227 excites
  6. L5:33 suppresses
  7. L6:2492 excites
  8. L10:319 excites
  9. L7:828 excites
  10. L11:453 excites
  11. L6:189 suppresses
  12. L5:56 excites
  13. L10:390 suppresses
  14. L6:1579 excites
  15. L9:132 excites
  16. L16:2983 excites
  17. L10:679 suppresses
  18. L6:150 excites
  19. L9:630 suppresses
  20. L5:1614 excites
  21. L16:847 suppresses
  22. L6:1841 excites
  23. L9:94 excites
  24. L14:1920 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 -1.51 and -1.452 values between -1.45 and -1.383 values between -1.38 and -1.324 values between -1.32 and -1.252 values between -1.25 and -1.193 values between -1.19 and -1.128 values between -1.12 and -1.067 values between -1.06 and -1.003 values between -1.00 and -0.938 values between -0.93 and -0.879 values between -0.87 and -0.8015 values between -0.80 and -0.7412 values between -0.74 and -0.6718 values between -0.67 and -0.6119 values between -0.61 and -0.5424 values between -0.54 and -0.4820 values between -0.48 and -0.4228 values between -0.42 and -0.3537 values between -0.35 and -0.2932 values between -0.29 and -0.2281 values between -0.22 and -0.16141 values between -0.16 and -0.09265 values between -0.09 and -0.03795 values between -0.03 and 0.04489 values between 0.04 and 0.10531 values between 0.10 and 0.16519 values between 0.16 and 0.23409 values between 0.23 and 0.29271 values between 0.29 and 0.3697 values between 0.36 and 0.4230 values between 0.42 and 0.4913 values between 0.49 and 0.553 values between 0.55 and 0.612 values between 0.61 and 0.682 values between 0.68 and 0.741 values between 0.74 and 0.812 values between 0.87 and 0.942 values between 1.00 and 1.06offnegativepositiveoften

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. MBED
  3. odi
  4. oris
  5. ###
  6. вар
  7. ount
  8. mientras
  9. correctamente
  10. hog

Suppresses

  1. #{@
  2. eren
  3. )?$
  4. RoutingModule
  5. ClickListener
  6. 了好多
  7. */ ⏎ ⏎ ⏎
  8. ' ⏎
  9. _ ⏎
  10. ⇆