Neuron Explorer

Qwen3-0.6B

action verbs in how-to steps

A neuron in layer 20 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 tends to fire on the action word in a step-by-step instruction, such as 'grow', 'twist', 'work' or 'cut'. It does not clearly change the topic when turned up.

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

    flower. An easier way to propagate Clivia miniata is by division, which can be done at any time of year. If you notice offsets (small plants growing from<|endoftext|>Geography: Midway Atoll

  2. #2 peak

    skunk is a mammal, with a unique striped black and white coat and defense system: ( glands beneath the tail that produce an oily, sulfurous substance which

  3. #3 peak

    the patients bone marrow as opposed to bone (painfully) harvested from the patient. The mixture is used as a "scaffold" for the new bone to grow around/inside.⏎Vision correction can be achieved using a semi-permantent contact

  4. #4 peak

    positioned⏎so you can fit U-shaped wire over the flat end and into the holes.⏎8. Slip large wire into the folded canvas slot of the net and twist the⏎9. Cut the twisted wire to no longer than 2" in

  5. #5 peak

    turn off the power at the service panel for the circuit that you will be connecting into before beginning work. Double check to make sure the circuit is actually safe to work on by using a noncontact voltage tester. In many cases, more than one

  6. #6 peak

    scrap paper, the longer the better.⏎begin with one small line⏎duplicate the segment 3 times, as explained earlier⏎leave the last fold folded, and here’s where our loop begins. start by counting

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 in the market of a new car and I have been getting it back from a previous owner. Now I have been in the current market of a car that has been been passed to her... I have been... been passed to

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 reign of Suleiman the Magnificent. However, after continued unrest in the interior, the Turks evacuated the region in 1630, leaving it in the hands of the Shiite Zaidi imams of Sanaa.

  2. #2 peak

    destroyed by the Turkish occupation forces. After the Turks left around 1770 the monastery was a ruin. From these ruins a new church was build. In 18

  3. #3 peak

    resulting in a four-storey keep, a three-towered forebuilding and impregnable defensive walls.⏎From 1216-17 the castle was besieged by the French,

  4. #4 peak

    extended to 70 years. When the period of copyright protection has ended, the written document, musical composition, book, picture, or other creative work is in the public domain. This means that no one owns the copyright and everyone is free

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. L14:1684 excites
  2. L16:2231 excites
  3. L18:2729 suppresses
  4. L19:1686 excites
  5. L11:1412 suppresses
  6. L16:1360 weakly excites
  7. L18:2195 weakly excites
  8. L14:2946 weakly excites
  9. L17:2673 weakly suppresses
  10. L14:1551 weakly excites
  11. L18:594 weakly excites
  12. L19:2064 weakly suppresses
  13. L13:2655 weakly excites
  14. L17:2977 weakly suppresses
  15. L13:2911 weakly excites
  16. L9:2467 weakly excites
  17. L16:2584 weakly excites
  18. L11:1299 weakly excites
  19. L16:2526 weakly suppresses
  20. L6:3045 weakly excites
  21. L19:796 weakly suppresses
  22. L6:2979 weakly suppresses
  23. L15:2975 weakly suppresses
  24. L18:1660 weakly excites

Feeds later layers

  1. L21:1723 suppresses
  2. L25:2021 suppresses
  3. L23:1734 weakly suppresses
  4. L21:1628 weakly excites
  5. L21:1022 weakly suppresses
  6. L21:375 weakly excites
  7. L26:791 weakly excites
  8. L22:1075 weakly excites
  9. L23:1747 weakly suppresses
  10. L21:1535 weakly suppresses 'Old' as part of a name or phrase
  11. L26:2047 weakly suppresses
  12. L21:981 weakly excites
  13. L23:2479 weakly suppresses
  14. L23:2980 weakly suppresses
  15. L26:1002 weakly suppresses
  16. L23:1894 weakly suppresses
  17. L25:1564 weakly suppresses word pieces before 'trans' or 'transfer'
  18. L23:1467 weakly excites
  19. L22:2776 weakly excites 'can' or 'could' before 'be found/read'
  20. L24:2293 weakly suppresses
  21. L22:1345 weakly excites
  22. L23:442 weakly suppresses
  23. L25:1723 weakly suppresses
  24. L21:1083 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.

1 values between -20.94 and -20.201 values between -20.20 and -19.471 values between -18.00 and -17.271 values between -14.33 and -13.591 values between -13.59 and -12.863 values between -12.12 and -11.391 values between -11.39 and -10.651 values between -10.65 and -9.922 values between -9.92 and -9.182 values between -9.18 and -8.454 values between -7.72 and -6.985 values between -6.98 and -6.253 values between -6.25 and -5.5111 values between -5.51 and -4.7813 values between -4.78 and -4.0416 values between -4.04 and -3.3148 values between -3.31 and -2.5765 values between -2.57 and -1.84104 values between -1.84 and -1.10414 values between -1.10 and -0.372519 values between -0.37 and 0.37443 values between 0.37 and 1.10130 values between 1.10 and 1.8354 values between 1.83 and 2.5732 values between 2.57 and 3.3014 values between 3.30 and 4.049 values between 4.04 and 4.775 values between 4.77 and 5.511 values between 5.51 and 6.243 values between 6.24 and 6.981 values between 7.71 and 8.45offnegativepositiveoften

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. ływ
  2. Indented
  3. 解决问题
  4. 삼
  5. ProductId
  6. cố
  7. 发力
  8. 对学生
  9. artykuł
  10. 米饭

Suppresses

  1. odelist
  2. donor
  3. 又被
  4. belonging
  5. 已被
  6. 遭
  7. donors
  8. redraw
  9. entieth
  10. blev