Neuron Explorer

Qwen3-0.6B

'con' starting a word piece

A neuron in layer 27 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 the syllable 'con' when it begins a word fragment, as in constrict, congregate, concomitant and condescended. Turning it down made the model output a stream of letter c's, and turning it up made it repeat words like 'finals' and 'whether'.

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

    brain’s reward center. It can also cause blood vessels to constrict and muscle relaxation. It may also impair memory and concentration. Nicotine can

  2. #2 peak

    a major, preventable cause of harm. To improve knowledge on anticoagulant safety, we plan to conduct bidirectional self-controlled case series studies, among concomitant users of warfarin and a skeletal muscle relaxant and, as

  3. #3 peak

    organic methods, with no chemical pesticides or fertilisers,” Crocco told IPS.⏎The crops planted in the school gardens, depending on the season, include lettuce, arugula, Swiss chard, spinach, bell peppers, tomatoes, squash,

  4. #4 peak

    every night, while others snore only occasionally.⏎What is sleep apnea?⏎Sleep apnea is similar to snoring, except that the airway becomes completely constricted. As a result, it becomes impossible to breathe. The typical pattern for

  5. #5 peak

    had a request to share more about that last one- so we’re meeting some snakes from around the world.⏎Today we’re meeting the black rat snake- a constrictor and helpful snake.⏎Black Rat Fun Facts:⏎- They are also known

  6. #6 peak

    not elective or historical. They reach right into those classrooms, affecting every single student. Students we worked with talked of being cat-called, harassed, objectified, condescended to and, most notably, excluded from the same spaces their male counterparts

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: Cc c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c

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

    known to result in appetite loss. Plus, the pain, fatigue and other<|endoftext|>This Super food is good source of fiber and has only 50 calories per one-cup serving⏎Pumpkin can keep you feeling full longer

  2. #2 peak

    ) and various other nutrients.⏎Protein content: 59% of calories. One cup (226 g) of cottage cheese with <|endoftext|>National Mammal Week

  3. #3 peak

    on fewer calories. Why? Because pumpkin is a good source of fiber and has only 50 calories per one-cup serving of mashed pumpkin. Recent research is even finding that pumpkin can help

  4. #4 peak

    oids, B vitamins, and fiber. It is also low in calories.⏎Nutritional Value and Health Effects⏎A 100-gram (3.5 oz) serving of oca provides almost 40 milligrams of vitamin C

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. L6:1963 weakly suppresses
  2. L3:576 weakly suppresses
  3. L4:2 weakly suppresses
  4. L26:3004 weakly suppresses
  5. L24:105 weakly suppresses
  6. L14:1627 weakly suppresses
  7. L7:102 weakly suppresses
  8. L3:59 weakly excites
  9. L21:237 weakly suppresses
  10. L0:1057 weakly excites
  11. L26:1564 weakly excites
  12. L23:206 weakly suppresses
  13. L18:1083 weakly suppresses
  14. L22:2365 weakly suppresses
  15. L26:581 weakly suppresses
  16. L15:2447 weakly suppresses
  17. L2:761 weakly excites
  18. L1:23 weakly suppresses
  19. L1:2677 weakly suppresses
  20. L14:2441 weakly suppresses
  21. L0:2701 weakly suppresses
  22. L21:236 weakly excites
  23. L16:1507 weakly excites
  24. L23:1685 weakly excites

Feeds later layers

None: this is the last layer.

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 -32.00 and -30.381 values between -28.76 and -27.152 values between -17.44 and -15.821 values between -12.58 and -10.972 values between -10.97 and -9.355 values between -9.35 and -7.732 values between -7.73 and -6.115 values between -6.11 and -4.507 values between -4.50 and -2.8814 values between -2.88 and -1.263583 values between -1.26 and 0.36269 values between 0.36 and 1.984 values between 1.98 and 3.605 values between 3.60 and 5.213 values between 6.83 and 8.451 values between 11.69 and 13.301 values between 13.30 and 14.921 values between 14.92 and 16.541 values between 31.10 and 32.72offnegativepositiveoften

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. 姊妹
  2. bih
  3. eses
  4. @stop
  5. 因为它
  6. crest
  7. _EDEFAULT
  8. Footer
  9. prises
  10. ⎦

Suppresses

  1. c
  2. c
  3. C
  4. C
  5. _c
  6. .c
  7. *c
  8. ;c
  9. cac
  10. →c