Neuron Explorer

Qwen3-0.6B

unclear: common core and function-docs terms

A neuron in layer 1 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron is hard to pin down. It fires a little on words like 'Standards' in 'Common Core Standards', and it dips on programming terms such as 'parameters', 'prototype' and 'value' in function explanations. Turning it up or down did not give a clear change in the writing.

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

    <|endoftext|>Thank you to DadsWorksheets.com for providing a VARIETY of math practice sheets! This website has been a great resource for practicing Math Common Core Standards. The Fact Family

  2. #2 peak

    une Function, 2017, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5707683/#:~:text=It%20is%20a%20potent%

  3. #3 peak

    formal construction and tone of a company job application.⏎This is the explanation as to why

  4. #4 peak

    history, and society.⏎About the authors⏎Elaine Treharne is Roberta Bowman Denning Professor of Humanities and Professor of English (and, by courtesy, of German Studies) at Stanford University, as well as Director of Stanford's

  5. #5 peak

    1: Download Drupal⏎First download the Drupal using following command:⏎# wget http://ftp.drupal.org/files/projects/drupal-7.33.zip⏎Step

  6. #6 peak

    of a lego brick⏎- Number of bricks needed = volume of planet Earth / volume of a lego brick⏎We will need to use the following two formulas:⏎You can now code this algorithm using Python to

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 been thinking about lately" (in Chinese) "我最近在思考着..." The user has been thinking about lately: "I have been thinking about lately" (in Chinese) "我最近在思考着..."

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

    by the function and can be a type C data.⏎- Function name. This is the identifier in which you can use to call your function.⏎- Argument. Parameters that are passed to function of any type of C data.⏎You will declare your

  2. #2 peak

    or above your loop function. You can always declare your function in two ways:⏎First and foremost, you will write part if a function is known as the function prototype above your loop function that consists of:⏎- Name of the function<|endoftext|>Taking time

  3. #3 peak

    E_INVALID | FE_DIVBYZERO | FE_OVERFLOW | FE_UNDERFLOW) is non-zero, an exception has been raised. An application should either examine the return value or check the floating point exception flags to detect exceptions.⏎An application can also set

  4. #4 peak

    footprint.⏎Advantages and disadvantages of using Sunpower solar panels⏎The use of solar energy is becoming

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. L0:178 suppresses
  2. L0:1224 suppresses
  3. L0:1301 suppresses
  4. L0:821 excites
  5. L0:9 suppresses
  6. L0:2742 weakly excites
  7. L0:1001 weakly suppresses
  8. L0:2185 weakly excites
  9. L0:25 weakly excites
  10. L0:106 weakly suppresses unclear: 'if' opening a statement, 'local'
  11. L0:156 weakly excites
  12. L0:1 weakly suppresses
  13. L0:177 weakly excites
  14. L0:2346 weakly excites
  15. L0:1712 weakly excites
  16. L0:463 weakly excites
  17. L0:561 weakly suppresses
  18. L0:1477 weakly excites
  19. L0:1591 weakly suppresses
  20. L0:1073 weakly suppresses
  21. L0:2296 weakly excites
  22. L0:97 weakly excites
  23. L0:2076 weakly excites
  24. L0:603 weakly excites

Feeds later layers

  1. L22:1915 excites
  2. L2:1637 suppresses
  3. L15:1277 weakly suppresses
  4. L5:1079 weakly suppresses
  5. L19:461 weakly suppresses
  6. L20:190 weakly excites
  7. L17:2149 weakly excites
  8. L3:30 weakly suppresses
  9. L19:288 weakly excites
  10. L7:921 weakly excites
  11. L12:2394 weakly excites
  12. L3:2715 weakly excites
  13. L6:1308 weakly excites
  14. L20:72 weakly excites
  15. L19:403 weakly excites
  16. L15:699 weakly suppresses
  17. L14:1926 weakly excites
  18. L16:823 weakly suppresses
  19. L24:883 weakly excites
  20. L11:1700 weakly excites
  21. L9:2523 weakly excites
  22. L21:2723 weakly suppresses
  23. L13:1400 weakly suppresses "years" after a distance or time number
  24. L24:1366 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.

4 values between -0.30 and -0.282 values between -0.28 and -0.264 values between -0.26 and -0.248 values between -0.24 and -0.2213 values between -0.22 and -0.2017 values between -0.20 and -0.1817 values between -0.18 and -0.1523 values between -0.15 and -0.1340 values between -0.13 and -0.1180 values between -0.11 and -0.09101 values between -0.09 and -0.07173 values between -0.07 and -0.05283 values between -0.05 and -0.03480 values between -0.03 and -0.01756 values between -0.01 and 0.02627 values between 0.02 and 0.04405 values between 0.04 and 0.06286 values between 0.06 and 0.08181 values between 0.08 and 0.10140 values between 0.10 and 0.1293 values between 0.12 and 0.1455 values between 0.14 and 0.1734 values between 0.17 and 0.1923 values between 0.19 and 0.2116 values between 0.21 and 0.2311 values between 0.23 and 0.2515 values between 0.25 and 0.276 values between 0.27 and 0.297 values between 0.29 and 0.323 values between 0.32 and 0.342 values between 0.34 and 0.361 values between 0.38 and 0.401 values between 0.49 and 0.511 values between 0.53 and 0.55offnegativepositiveoften

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. gne
  3. endas
  4. desn
  5. scn
  6. дум
  7. bucks
  8. $"
  9. иона
  10. insula

Suppresses

  1. 有限责任
  2. mest
  3. 成年人
  4. forgetting
  5. memory
  6. зап
  7. 骨干
  8. 特长
  9. 同事们
  10. conting