Neuron Explorer

Qwen3-0.6B

'at' in 'not at all' and 'at all'

A neuron in layer 21 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 'at' in the phrase 'not at all' or 'at all', as in 'there aren't any trees at all'. Turning it up made the model write 'all at all' over and over.

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

    ; it could rise, fall, and rise again, as in a W. (Or, in the direst situation, the recovery may not be a recovery at all, but just an L-shaped continuation of economic misery.)⏎But not everyone subscri

  2. #2 peak

    the water and waste in the system to flow smoothly out to the sewer or septic tank. If there were no vent, the system would drain slowly or not at all, not unlike a water-filled straw when you seal one end with your finger

  3. #3 peak

    s" the others, it masks them more or less: in other words, the other sound are perceived as much weaker than they actually are, or even not at all. This phenomenon is of course taken into account by all the audio systems of

  4. #4 peak

    percent reported eating any meat, which historically was a diet staple, and many of them had only had one meal in 24 hours. Some had not eaten at all.⏎The Central African Republic has the highest level of hunger in the world,

  5. #5 peak

    influence how substances are metabolized and experienced. Such biological differences can influence whether substance use continues and increases, remains “social” and occasional, or does not occur at all. In the early stages, however, those more genetically vulnerable to addiction are

  6. #6 peak

    had only 18 percent marketable ears. This wasn't much better than the 6 percent marketable ears produced in the plots that received no sprays at all."⏎The authors predict that growers could realize increased profits with Bt sweet corn

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 have a 3D map, and I have a 2D grid where I can move, but I also have another, more open, and, more, than once, with the same number of lines and columns. Now,

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

    tragic existence, blazing a trail of death and destruction that prompted mobs of angry villagers to take up torches and pitchforks against him on the silver screen. Never once during his rampage, however, did the monster question the honesty of his ultimate

  2. #2 peak

    won’t arrive at the Jezero Crater until February 18, 2021. But if the helicopter soars, it’ll provide never-before-seen views of the Red Planet.⏎Popayán Colombia https://www

  3. #3 peak

    6th Field Artillery bands played at the monument's dedicationon July 3, 1929. By 1949,<|endoftext|>Never underestimate the power of a paper towel roll.⏎Just the other day my five-year

  4. #4 peak

    , thus increasing the resistence. The microcontroller senses this and turns off a few of the LEDs as a result. Make one of these guys and you’ll never again have to

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. L18:569 excites
  2. L20:1922 suppresses
  3. L16:752 excites
  4. L10:111 suppresses
  5. L20:2853 suppresses
  6. L10:39 excites
  7. L19:2184 suppresses
  8. L8:7 excites
  9. L18:2553 suppresses
  10. L18:1933 excites
  11. L5:158 weakly suppresses
  12. L7:49 weakly suppresses
  13. L17:1778 weakly suppresses
  14. L6:13 weakly excites
  15. L18:1763 weakly excites comma in 'not X, but Y' contrasts
  16. L6:2339 weakly suppresses "as" completing an "as ... as" comparison
  17. L17:1613 weakly excites
  18. L17:1644 weakly suppresses
  19. L7:662 weakly excites
  20. L18:178 weakly suppresses
  21. L5:28 weakly excites
  22. L7:110 weakly suppresses
  23. L10:386 weakly excites
  24. L15:386 weakly suppresses

Feeds later layers

  1. L27:599 strongly excites
  2. L26:835 strongly excites
  3. L24:1252 strongly excites
  4. L24:856 excites
  5. L25:501 suppresses
  6. L26:840 suppresses
  7. L26:1210 excites
  8. L26:2537 suppresses
  9. L23:1458 excites
  10. L27:769 weakly excites
  11. L23:1305 weakly excites
  12. L26:1800 weakly excites
  13. L25:1241 weakly excites
  14. L23:2703 weakly suppresses
  15. L22:1524 weakly excites
  16. L23:2966 weakly suppresses
  17. L27:1119 weakly excites
  18. L23:2095 weakly suppresses
  19. L27:1315 weakly excites
  20. L25:582 weakly suppresses
  21. L24:380 weakly suppresses the word 'at' in place phrases
  22. L23:2671 weakly excites
  23. L22:2641 weakly suppresses
  24. L26:684 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.

2 values between -9.96 and -9.341 values between -8.10 and -7.481 values between -7.48 and -6.861 values between -6.24 and -5.621 values between -5.62 and -5.001 values between -5.00 and -4.382 values between -3.76 and -3.144 values between -3.14 and -2.524 values between -2.52 and -1.898 values between -1.89 and -1.2762 values between -1.27 and -0.651582 values between -0.65 and -0.032186 values between -0.03 and 0.5929 values between 0.59 and 1.218 values between 1.21 and 1.833 values between 1.83 and 2.452 values between 2.45 and 3.073 values between 3.07 and 3.691 values between 3.69 and 4.311 values between 4.31 and 4.931 values between 6.17 and 6.791 values between 8.03 and 8.651 values between 9.28 and 9.902 values between 9.90 and 10.521 values between 14.24 and 14.86offnegativepositiveoften

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. at
  2. At
  3. _at
  4. At
  5. at
  6. -at
  7. →at
  8. .at
  9. (at
  10. /at

Suppresses

  1. .za
  2. INCT
  3. ЛА
  4. hra
  5. PHA
  6. PSA
  7. стал
  8. cratch
  9. põ
  10. mdir