Neuron Explorer

Qwen3-0.6B

the word 'average' before a quantity

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 lights up on the word 'average', especially in phrases like 'an average of 10 percent' or 'an average teacher'. When it was turned up, the model kept writing about averages and computing them.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    group.⏎The results of the studied showed that the groups with the most weight loss were the diet-only and diet and exercise group with women in those groups losing an average of 10 percent of their starting weight, according to the study.⏎Additionally

  2. #2 peak

    year.⏎Salary by Employment Sector⏎As of 2011, the highest-paid electricians were those who taught at business schools. These workers earned an average of $79,250 a year. Most electricians worked as

  3. #3 peak

    data were derived from national food consumption surveys, including more than 1000 individuals by country (women and men/18-64 years). Average consumption, GHGE (gCO2eq) and nutritional composition of 1

  4. #4 peak

    and weeding torches calories, too. A 154-pound person burns an average of 330 calories per hour through gardening, according to the US Centers

  5. #5 peak

    point) they hosted. Overall, 57.9% of trees in the study sites carried at least one liana. The infested trees hosted an average of 1.

  6. #6 peak

    of effectiveness) each year produces over $350,000 more income for his/her students compared to an average teacher. On the the other hand, a poor teacher (one in the bottom

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: the problem is the same as the set of possible ones? So, what's the problem with the set of possible ones? Let's say that the problem is the same as the set of possible ones? So, what's the problem?

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

    :⏎Top Suggestions Vocabulary Worksheet Mix :⏎Vocabulary Worksheet Mix Whether you re planning a summer lesson or

  2. #2 peak

    tools and materials of high quality came into use, masterpieces of calligraphic art were produced, e.g., the Irish Book of Kells (8th cent.; Trinity College, Dublin; see under Ceanannus Mór) and

  3. #3 peak

    has been consistently practiced as a major aesthetic expression. In China, from the 5th cent. B.C., when it was first used, calligraphy has always been considered

  4. #4 peak

    22, at 888-835-5322 (TTY) or at email@example.com.<|endoftext|>Table of Contents :⏎Top Suggestions Senior Kindergarten Writting Worksheet :⏎Senior Kindergarten Writting Worksheet In

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. L3:2163 suppresses
  2. L2:2862 excites
  3. L3:2897 excites
  4. L3:2291 suppresses
  5. L3:2686 weakly excites
  6. L2:356 weakly excites
  7. L2:2692 weakly excites
  8. L1:1464 weakly excites
  9. L3:50 weakly excites
  10. L1:2801 weakly excites
  11. L0:2849 weakly excites
  12. L2:84 weakly excites
  13. L3:421 weakly suppresses
  14. L3:3054 weakly suppresses
  15. L1:930 weakly excites
  16. L3:1755 weakly suppresses
  17. L3:2875 weakly suppresses
  18. L3:115 weakly suppresses
  19. L3:272 weakly excites
  20. L3:2047 weakly excites
  21. L3:59 weakly excites
  22. L2:1657 weakly excites
  23. L0:1301 weakly suppresses
  24. L1:2683 weakly excites

Feeds later layers

  1. L16:2836 excites
  2. L6:957 excites
  3. L8:1463 excites
  4. L21:1242 excites
  5. L7:2826 excites
  6. L16:1623 excites
  7. L8:2723 weakly excites
  8. L8:2150 weakly excites
  9. L11:1534 weakly excites
  10. L7:795 weakly suppresses
  11. L6:1153 weakly excites
  12. L5:1084 weakly suppresses
  13. L5:1656 weakly excites 'of' after end, top or bottom
  14. L8:910 weakly suppresses
  15. L16:1810 weakly excites
  16. L11:2115 weakly suppresses
  17. L22:2977 weakly excites
  18. L11:1160 weakly excites
  19. L6:674 weakly suppresses comma after a 'said' or 'answered' verb
  20. L10:503 weakly excites
  21. L5:811 weakly suppresses
  22. L8:2846 weakly suppresses
  23. L6:2961 weakly suppresses
  24. L17:1445 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 -0.71 and -0.5514 values between -0.55 and -0.39115 values between -0.39 and -0.23914 values between -0.23 and -0.082615 values between -0.08 and 0.08226 values between 0.08 and 0.2421 values between 0.24 and 0.401 values between 0.87 and 1.021 values between 5.43 and 5.58offnegativepositiveoften

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. emme
  2. ]->
  3. -sama
  4. point
  5. Москве
  6. 散布
  7. embro
  8. ]][
  9. ensity
  10. ocus

Suppresses

  1. zá
  2. ivers
  3. 不完
  4. Patriot
  5. KR
  6. ران
  7. ponible
  8. 国王
  9. idon
  10. ват