Neuron Explorer

Qwen3-0.6B

'in' as a phrase particle

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 responds to the word 'in' when it finishes a phrase, like 'take things in hand', 'went up in the air', or 'close in late June'. The matches are weaker than the top neurons on this page. Turning it up did not produce related writing.

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

    for getting regulations published in the Federal Register, Rumsfeld decided to take things in hand. He assembled Jack Grayson, the chairman of the Price Commission, and

  2. #2 peak

    neighborhood of Seattle and is a University of Washington graduate. During WWII his parents were sent to the Minidoka concentration camp in Idaho.⏎A teenager is pulled back in time to witness her grandmother’s experiences in World War II-era Japanese internment camps

  3. #3 peak

    source is also<|endoftext|>Since 2002, when HAVA was enacted, computer scientists have released over three dozen studies of the software-based voting systems rushed in by HAVA. Every one of these studies has concluded that software-based optical scanners

  4. #4 peak

    check Fairness Act would close loopholes, strengthen business incentives to end pay discrimination, prohibit retaliation against workers who share wage information, and bring the Equal Pay Act in line with other civil rights laws. Passing this act does not just affect women,

  5. #5 peak

    men heard the shells edging closer. “And then bang!—one of the big shells struck the ground right in front of us and we all went up in the air. But we all come down again.”⏎York was showered with honors

  6. #6 peak

    great migration,” more societies are becoming like the United States, with a number of diaspora communities. But diasporas not only change the societies that took them in, they can also impact the countries they left behind.⏎“There’s a tendency to

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: A problem where a grid has a certain number of rows, and columns, and there's a certain number of rows? Let me first understand the context of the problem. This problem, the user's question, the grid has a certain number

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

    In the real world, speakers sometimes also revert to illeism when they want to create

  2. #2 peak

    In the training room, athletes and athletic trainers often use the application of moist heat to boost

  3. #3 peak

    In a capitalist society, all human relationships are voluntary. Men are free to cooperate or not

  4. #4 peak

    ?⏎In a common cathode RGB LED, the cathode of the internal LEDs are all connected

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. L3:493 suppresses
  2. L2:2307 excites
  3. L1:1179 suppresses 'in' after 'result' (results in)
  4. L0:1280 weakly excites
  5. L0:1161 weakly excites
  6. L2:2918 weakly suppresses
  7. L1:290 weakly excites
  8. L3:1756 weakly suppresses
  9. L1:1718 weakly excites
  10. L0:536 weakly excites
  11. L1:2201 weakly excites
  12. L3:1116 weakly excites
  13. L0:2143 weakly excites
  14. L2:207 weakly suppresses
  15. L3:2739 weakly suppresses
  16. L0:1432 weakly excites
  17. L2:1541 weakly suppresses
  18. L1:1037 weakly suppresses
  19. L3:621 weakly suppresses
  20. L1:2953 weakly excites
  21. L2:1034 weakly suppresses
  22. L3:2871 weakly excites
  23. L2:1899 weakly excites
  24. L3:159 weakly excites

Feeds later layers

  1. L7:2878 weakly suppresses
  2. L10:2604 weakly suppresses
  3. L20:1356 weakly suppresses
  4. L18:1372 weakly suppresses
  5. L6:823 weakly suppresses
  6. L14:2709 weakly excites
  7. L12:1808 weakly suppresses
  8. L11:386 weakly suppresses
  9. L11:2172 weakly suppresses
  10. L15:2102 weakly excites
  11. L6:2072 weakly excites
  12. L26:1102 weakly suppresses fibromyalgia and hour counts
  13. L17:272 weakly suppresses
  14. L8:2722 weakly suppresses
  15. L6:2844 weakly suppresses
  16. L5:199 weakly suppresses
  17. L26:1469 weakly suppresses
  18. L18:1817 weakly excites
  19. L16:623 weakly suppresses
  20. L8:2610 weakly excites
  21. L17:612 weakly suppresses
  22. L23:2080 weakly suppresses
  23. L21:2501 weakly suppresses
  24. L12:1518 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 -2.48 and -2.342 values between -2.34 and -2.191 values between -2.19 and -2.052 values between -1.76 and -1.622 values between -1.47 and -1.331 values between -1.33 and -1.193 values between -1.19 and -1.043 values between -1.04 and -0.903 values between -0.90 and -0.755 values between -0.75 and -0.617 values between -0.61 and -0.4758 values between -0.47 and -0.32315 values between -0.32 and -0.181339 values between -0.18 and -0.041759 values between -0.04 and 0.11294 values between 0.11 and 0.2548 values between 0.25 and 0.4018 values between 0.40 and 0.545 values between 0.54 and 0.686 values between 0.68 and 0.837 values between 0.83 and 0.973 values between 0.97 and 1.112 values between 1.11 and 1.265 values between 1.26 and 1.404 values between 1.40 and 1.552 values between 1.55 and 1.692 values between 1.69 and 1.834 values between 1.83 and 1.981 values between 2.12 and 2.261 values between 2.26 and 2.412 values between 2.41 and 2.551 values between 2.55 and 2.702 values between 3.13 and 3.27offnegativepositiveoften

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. imates
  2. esub
  3. 发展机遇
  4. 覆
  5. 收购
  6. trends
  7. -gen
  8. (“
  9. preserve
  10. ).</

Suppresses

  1. ulado
  2. ยาว
  3. bras
  4. flater
  5. _different
  6. ilitation
  7. unfinished
  8. assoc
  9. measurement
  10. ?