Neuron Explorer

Qwen3-0.6B

the word 'whole'

A neuron in layer 25 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 "whole" in phrases like "whole number", "whole grains" or "whole egg". Turning it up made the model repeat the word "number" over and over, which fits the word it tends to predict next.

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

    to the nearest whole number.)<|endoftext|>Most of the sounds Sperm Whales produce are clicks, from

  2. #2 peak

    to reach their full potential by passively listening to the teacher and filling in worksheets they must engage in This is called a mixed number which is a mixture of a whole bbc bitesize to add the two fractions together print out the worksheet below from twink

  3. #3 peak

    for Americans recommends that at least half of all the grains you eat are whole grains. Chances are you eat lots of grains already. But are they whole

  4. #4 peak

    for your dataset may be different.⏎The stem-and-leaf plot allows us to see where data may be aggregated.⏎Use the app above to select a dataset with whole numbers. The app will output the dataset's stem-and-leaf plot once Submit

  5. #5 peak

    folks who have high blood pressure, even more evidence points to the benefits of the DASH diet (an eating plan that encourages a lot of fruits, vegetables, whole grains, and heart-healthy fats). Since these diseases and conditions are all too

  6. #6 peak

    one larger whole.⏎As it becomes more difficult to shrink transistors in size, and impr

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 best way to fit together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together together

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

    ucius’ philosophy revolved around two concepts: the nobleman and the establishment of a well-ordered society. The nobleman (君子, pinyin: jū

  2. #2 peak

    no leader since the priest or priestess read the will of deities.⏎Leadership, Odin, messenger birds, Lugh Long-arm, Odin, Tyr, Amled saga, neck-ring, Cernunnos, the Stag, Trinity

  3. #3 peak

    Perseids meteor shower.⏎There are many meteor showers during the year, but the Perseids in August, the Leonids in November, and the Geminids in December are predict

  4. #4 peak

    largest hang like many-limbed chandeliers, turning slowly in the current. When I get closer, I can see the individual polyps, the anemone-like creatures that make up each fragmented colony. Nestled in the cren

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. L18:242 weakly suppresses
  2. L11:1091 weakly excites
  3. L23:270 weakly excites
  4. L9:1795 weakly suppresses
  5. L2:2863 weakly suppresses
  6. L24:105 weakly suppresses
  7. L16:819 weakly excites
  8. L1:987 weakly excites
  9. L13:2877 weakly suppresses
  10. L3:1939 weakly suppresses
  11. L17:1061 weakly suppresses
  12. L3:62 weakly suppresses
  13. L20:183 weakly excites
  14. L2:343 weakly suppresses
  15. L3:508 weakly excites the final zero of decades like '1980s'
  16. L0:2571 weakly suppresses
  17. L10:933 weakly suppresses
  18. L5:2466 weakly excites
  19. L4:1323 weakly suppresses
  20. L16:11 weakly suppresses
  21. L22:2003 weakly suppresses
  22. L10:190 weakly suppresses
  23. L11:2637 weakly suppresses
  24. L7:846 weakly excites

Feeds later layers

  1. L26:861 excites
  2. L27:974 excites
  3. L27:1479 weakly excites
  4. L26:2757 weakly excites
  5. L27:2145 weakly suppresses
  6. L26:880 weakly suppresses
  7. L27:572 weakly suppresses
  8. L26:2628 weakly excites
  9. L27:1568 weakly excites
  10. L26:1618 weakly excites
  11. L27:886 weakly excites
  12. L27:2484 weakly suppresses
  13. L27:1389 weakly suppresses
  14. L26:748 weakly suppresses
  15. L27:1812 weakly suppresses
  16. L27:229 weakly suppresses
  17. L27:1631 weakly suppresses
  18. L27:2910 weakly suppresses
  19. L27:2795 weakly suppresses
  20. L26:969 weakly excites
  21. L27:1944 weakly excites
  22. L27:1754 weakly suppresses
  23. L26:1520 weakly excites
  24. L27:634 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 -12.38 and -11.421 values between -9.49 and -8.531 values between -7.56 and -6.604 values between -5.63 and -4.673 values between -4.67 and -3.703 values between -3.70 and -2.7410 values between -2.74 and -1.7839 values between -1.78 and -0.812827 values between -0.81 and 0.15969 values between 0.15 and 1.1229 values between 1.12 and 2.084 values between 2.08 and 3.056 values between 3.05 and 4.014 values between 4.01 and 4.971 values between 6.90 and 7.871 values between 8.83 and 9.801 values between 9.80 and 10.761 values between 10.76 and 11.721 values between 12.69 and 13.651 values between 14.62 and 15.581 values between 25.22 and 26.19offnegativepositiveoften

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. 数
  2. number
  3. number
  4. 數
  5. numbers
  6. (number
  7. số
  8. Number
  9. -number
  10. num

Suppresses

  1. iggs
  2. @Json
  3. Frederick
  4. EDURE
  5. opause
  6. chaft
  7. Inflate
  8. QGraphics
  9. iedad
  10. FITNESS