Neuron Explorer

Qwen3-0.6B

advice words like avoid, limit, recommend

A neuron in layer 13 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on advice words such as 'avoid', 'limit' and 'recommend', especially in phrases like 'Foods to avoid' or 'It's best to avoid'. Turning it up did not bring out that advice theme; the model wrote repetitive or unrelated text.

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

    .⏎But like I said before: a ferret cannot digest plant matter which means that giving them anything other than meat will lead to health problems.⏎Foods to avoid:⏎-

  2. #2 peak

    out for sure ingredients. Here's an inventory of toxic foods you would like to avoid.⏎- Chocolate is one among a lot of renowned toxic human food a dog will

  3. #3 peak

    ings.⏎4. Limit sugary and acidic foods or drinks:⏎Sugary and acidic foods and beverages can erode your tooth enamel and cause cavities. Limit your intake of sugary or acidic foods and beverages, and drink plenty of water

  4. #4 peak

    would like to support the mission of Open Culture, consider making a donation to our site. It’s hard to rely 100% on ads, and your contributions will help us continue providing the best free cultural and educational materials to learners everywhere

  5. #5 peak

    the fitness industry. As does the “no pain, no gain” myth, which came to prominence in the early 1980

  6. #6 peak

    tires and who overdrive the conditions.<|endoftext|>Viewing page 28 of 101⏎It looks like you're using a mobile device. We recommend using a physical keyboard for transcription entry.⏎28 REPORT NATIONAL ADVISORY COMMIT

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 have been thinking about the things I have been thinking about the things I have been thinking about the things I have been thinking about. I have been thinking about the things I have been thinking about the things I have been thinking about the things

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

    presents both an opportunity and a challenge for the planet. Carlos Barria / Reuters⏎A nurse takes care of newborn

  2. #2 peak

    made their way through the “land,” they count their tokens and see if they made it through correctly by solving another addition problem at the end. This is a great self-checking component of the activity.⏎SUBTRACTION FACT LAND HELPS STUD

  3. #3 peak

    u-substitution twice (second time with w)⏎⇐ Use this menu to view and help create subtitles for this video in many different languages. You'll probably want to hide YouTube's captions if using these subtitles.⏎- Let's see if we can

  4. #4 peak

    part of your educational outreach around Earth Day, or use it as part of your lesson—it’s a great poster to hang in the classroom! We encourage you to help us spread this important

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. L11:592 suppresses
  2. L10:750 excites
  3. L12:517 suppresses
  4. L12:833 suppresses
  5. L8:1754 excites
  6. L10:203 excites
  7. L8:1822 suppresses
  8. L12:210 excites
  9. L10:1447 excites
  10. L11:1117 suppresses
  11. L10:920 excites
  12. L12:332 excites
  13. L12:967 suppresses
  14. L12:889 weakly suppresses
  15. L12:1232 weakly excites
  16. L8:1126 weakly excites unclear: scattered technical nouns
  17. L12:2389 weakly excites
  18. L10:641 weakly excites
  19. L11:782 weakly suppresses
  20. L11:431 weakly suppresses
  21. L12:1668 weakly excites
  22. L11:1074 weakly suppresses
  23. L12:2411 weakly excites
  24. L12:1136 weakly excites

Feeds later layers

  1. L14:2194 suppresses
  2. L16:669 excites
  3. L15:2295 suppresses
  4. L15:2748 suppresses
  5. L16:2813 suppresses
  6. L14:188 suppresses
  7. L16:2346 suppresses
  8. L16:1886 suppresses
  9. L16:2528 suppresses
  10. L15:1450 suppresses
  11. L14:924 excites
  12. L14:1372 excites
  13. L14:3028 suppresses
  14. L18:413 suppresses
  15. L14:1288 excites
  16. L14:121 suppresses
  17. L16:1225 excites
  18. L14:2183 suppresses
  19. L15:2409 suppresses
  20. L14:1339 excites
  21. L14:197 suppresses
  22. L18:412 suppresses
  23. L16:1587 excites
  24. L15:125 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 -2.67 and -2.541 values between -2.42 and -2.301 values between -2.06 and -1.942 values between -1.69 and -1.572 values between -1.57 and -1.451 values between -1.45 and -1.337 values between -1.33 and -1.219 values between -1.21 and -1.0818 values between -1.08 and -0.9619 values between -0.96 and -0.8438 values between -0.84 and -0.7253 values between -0.72 and -0.6079 values between -0.60 and -0.48132 values between -0.48 and -0.35235 values between -0.35 and -0.23422 values between -0.23 and -0.111538 values between -0.11 and 0.01912 values between 0.01 and 0.13232 values between 0.13 and 0.2590 values between 0.25 and 0.3845 values between 0.38 and 0.5030 values between 0.50 and 0.6215 values between 0.62 and 0.7410 values between 0.74 and 0.865 values between 0.86 and 0.982 values between 0.98 and 1.111 values between 1.11 and 1.231 values between 1.23 and 1.352 values between 1.47 and 1.591 values between 1.59 and 1.711 values between 1.71 and 1.841 values between 1.84 and 1.961 values between 1.96 and 2.081 values between 2.08 and 2.20offnegativepositiveoften

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. ков
  2. ialis
  3. 地说
  4. [at
  5. olarity
  6. ności
  7. psychologists
  8. 듬
  9. RefCount
  10. strup

Suppresses

  1. antry
  2. aha
  3. ede
  4. deen
  5. 莺
  6. anceled
  7. edith
  8. ah
  9. الحوثي
  10. arbon