Neuron Explorer

Qwen3-0.6B

'case' after 'in this'

A neuron in layer 9 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 lights up on the word 'case' in phrases like 'in this case', and on words that point back to the current example, such as 'here'. Turning it up made the model write about examples and the question at hand, though loosely.

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

    you plural|⏎A direct object is a noun following the verb that answers the questions what? or whom? Remember a pronoun replaces a noun, and in this case the noun is a direct object. Verbs that have a direct object are called

  2. #2 peak

    Hawaii (figure 2). When hydroacoustic activity has been of sufficient amplitude and duration, it has often been inferred to suggest submarine eruptions. In this case the signals were relatively weak and of modest duration, indicating seismic swarms without a

  3. #3 peak

    to the mites. The upshot is that a rash results.⏎So there are two ways to approach the problem: treat the underlying disease — CLL in your case

  4. #4 peak

    recently passed separate laws in 2001 that<|endoftext|>Pulse Width Modulation (or PWM) is a technique for controlling power. We also use it here to control the brightness of each of the LEDs. … Roughly every

  5. #5 peak

    When we judge an action as morally right or wrong, we rely on our capacity to infer the actor's mental states (e.g., beliefs, intentions). Here, we test the hypothesis that the right

  6. #6 peak

    in Tucson,⏎and lead author of the study, which appears in today's Nature.⏎"Normally, you have the greatest thickness underneath the tallest mountains⏎in this case

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 can be an animal of the year." So, how many times is the word used in a year?" Let me explain the word in a year, and what is its value in a year." So, how many times is the

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

    is the most appropriate management?⏎vitamin D and calcium levels⏎MRI of the pelvis⏎long leg cast⏎aspiration of the knee⏎A 3-year-old male presents with inability to bear weight on his right leg for the past

  2. #2 peak

    answers is unusual, for students scoring the way John and Barbara scored. If it is unusual, how unusual is it? If it would only occur one time in a million (or even less often) by chance alone, the instructor may conclude that

  3. #3 peak

    problems: one each of addition, subtraction, multiplication, division, fractions, and percents. I stressed "grade-appropriate" because single digit addition, for example, is not something most 5th graders need to practice. My expectations

  4. #4 peak

    seems appropriate to have some humility around this energy transition,” said co-author David Rapson, an associate professor at the UC Davis Economics Department. “Approaches that leave multiple technology pathways open are

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. L7:726 suppresses
  2. L8:401 suppresses
  3. L7:34 excites
  4. L8:355 excites
  5. L7:2291 suppresses
  6. L8:543 weakly excites "Since" starting a reason clause
  7. L7:318 weakly suppresses
  8. L7:267 weakly suppresses
  9. L2:1385 weakly suppresses
  10. L8:824 weakly suppresses
  11. L8:374 weakly suppresses
  12. L5:178 weakly suppresses
  13. L8:330 weakly excites
  14. L5:396 weakly suppresses
  15. L1:2372 weakly suppresses
  16. L8:207 weakly suppresses
  17. L3:487 weakly suppresses
  18. L7:1106 weakly suppresses
  19. L8:1884 weakly suppresses
  20. L8:606 weakly excites
  21. L6:2135 weakly suppresses
  22. L7:1418 weakly suppresses
  23. L1:1395 weakly suppresses
  24. L2:1646 weakly suppresses

Feeds later layers

  1. L12:2575 excites
  2. L12:976 suppresses
  3. L12:1602 excites
  4. L10:805 suppresses
  5. L12:1002 excites
  6. L16:2499 suppresses
  7. L11:549 weakly suppresses
  8. L15:1221 weakly excites
  9. L22:1854 weakly excites
  10. L14:2718 weakly excites
  11. L11:78 weakly suppresses
  12. L11:2619 weakly suppresses
  13. L15:86 weakly excites
  14. L11:54 weakly excites
  15. L12:2942 weakly excites
  16. L13:669 weakly excites
  17. L15:753 weakly suppresses
  18. L11:994 weakly suppresses
  19. L10:1036 weakly excites
  20. L13:320 weakly excites
  21. L10:996 weakly excites
  22. L10:961 weakly suppresses
  23. L14:2848 weakly excites
  24. L11:985 weakly 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 -1.32 and -1.151 values between -1.15 and -0.981 values between -0.98 and -0.802 values between -0.80 and -0.6310 values between -0.63 and -0.4636 values between -0.46 and -0.29367 values between -0.29 and -0.122342 values between -0.12 and 0.06950 values between 0.06 and 0.23162 values between 0.23 and 0.4027 values between 0.40 and 0.574 values between 0.57 and 0.751 values between 0.75 and 0.921 values between 0.92 and 1.091 values between 2.47 and 2.641 values between 2.64 and 2.811 values between 5.40 and 5.57offnegativepositiveoften

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. case
  2. cased
  3. case
  4. onaut
  5. ocoa
  6. isma
  7. .case
  8. -task
  9. itate
  10. хот

Suppresses

  1. }*/⏎
  2. _under
  3. 綦
  4. Portugal
  5. UNDER
  6. цип
  7. }*/⏎
  8. 又被
  9. vej
  10. 韬