Neuron Explorer

Qwen3-0.6B

dollar sign before amounts

A neuron in layer 23 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 dollar sign that comes right before a price or amount of money, like "$0.05" or "$12,000". When it was turned up, the written text drifted into odd, unrelated rambling, so the steering results don't show a clear theme.

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

    ), the nickel ($0.05), the dime ($0.10), the quarter ($

  2. #2 peak

    4-5 hours a day which are too warm for cow comfort. The estimated impact on milk production is 6 kg MS per cow over summer, or $12,000 for a 300-cow herd at

  3. #3 peak

    , tested the strength of common building materials such as pipes, and above-ground structures such as bridges and power lines. During the first test run at Cornell University’s $2.1 million earthquake lab, researchers simulated the effects of an earthquake on a

  4. #4 peak

    first to come up with these sorts of self-reliant devices. Back in 2015, for example, a separate team developed a tiny, $0.20 temperature sensor that didn’t require a battery or constant source of

  5. #5 peak

    defined, however some architectural considerations can be addressed and some conclusions can be reached.⏎Structure of Information⏎Consider these structures of data that can aid learning.⏎- The $M$ features of user linguistic response can be saved within a user session as a

  6. #6 peak

    , almost $8 for a gallon of gasoline and $6.50 for a gallon of

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: it's not a birth ritual in the sense that it's not in the birth pattern and not in the birth ritual in the ritual. It's not a birth in the birth ritual, it's not a birth in the birth pattern. It

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

    slash ( `` ). Some common escape sequences are: - `n` Newline - Prints a newline character - `t ` Tab - Prints a tab character - `\ ` Backslash - Prints a backslash ( \ ) - `' ` Single quote

  2. #2 peak

    *) memchr (str, 'p', strlen(str));⏎printf ("'p' found at position %d.\n", pch-str+1);⏎printf ("'p' not found.\

  3. #3 peak

    Whether you’re a<|endoftext|>- Enter a word for the dictionary definition.⏎From The Collaborative International Dictionary of English v.0.48:⏎Nectarine \Nec`tar*ine"\ (n

  4. #4 peak

    $\quad+\quad$ or $\quad-$<|endoftext|>Design for a Mosaic⏎Albert Joseph Moore

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. L19:329 strongly suppresses
  2. L22:326 excites
  3. L17:719 suppresses
  4. L21:1042 suppresses
  5. L21:588 excites
  6. L19:1040 suppresses
  7. L21:1374 excites
  8. L22:2028 excites
  9. L17:527 suppresses
  10. L18:724 suppresses
  11. L18:153 suppresses
  12. L21:2508 suppresses
  13. L20:934 weakly excites
  14. L8:16 weakly suppresses
  15. L20:594 weakly suppresses
  16. L19:58 weakly excites
  17. L2:28 weakly excites
  18. L3:204 weakly excites
  19. L18:1351 weakly suppresses
  20. L20:496 weakly excites
  21. L17:2329 weakly suppresses
  22. L22:2107 weakly excites
  23. L1:2340 weakly suppresses
  24. L18:569 weakly excites

Feeds later layers

  1. L26:473 strongly excites
  2. L26:245 excites
  3. L24:54 excites
  4. L25:391 excites
  5. L27:171 excites
  6. L27:430 excites
  7. L25:630 excites
  8. L25:560 excites
  9. L27:1024 excites
  10. L24:458 excites
  11. L27:1480 excites
  12. L24:218 excites
  13. L24:256 excites
  14. L25:524 excites
  15. L25:487 suppresses
  16. L25:276 excites
  17. L25:380 excites
  18. L25:1952 suppresses
  19. L24:117 weakly excites
  20. L24:1601 weakly excites
  21. L25:113 weakly excites
  22. L26:540 weakly suppresses
  23. L27:1241 weakly excites
  24. L24:653 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 -2.28 and -2.0511 values between -2.05 and -1.8320 values between -1.83 and -1.6070 values between -1.60 and -1.38156 values between -1.38 and -1.15207 values between -1.15 and -0.93296 values between -0.93 and -0.70478 values between -0.70 and -0.48892 values between -0.48 and -0.261592 values between -0.26 and -0.03161 values between -0.03 and 0.194 values between 0.42 and 0.642 values between 0.64 and 0.874 values between 0.87 and 1.093 values between 1.32 and 1.541 values between 1.54 and 1.771 values between 1.77 and 1.991 values between 1.99 and 2.222 values between 2.44 and 2.671 values between 2.67 and 2.891 values between 3.12 and 3.341 values between 4.46 and 4.691 values between 6.04 and 6.261 values between 6.49 and 6.71offnegativepositiveoften

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. \
  2. \%
  3. \v
  4. \t
  5. \d
  6. ^\
  7. \r
  8. \
  9. ^{
  10. (\

Suppresses

  1. อำนาจ
  2. ($("#
  3. =$("#
  4. �
  5. **)
  6. 在现场
  7. ="--
  8. impres
  9. ($(".
  10. (MouseEvent