Neuron Explorer

Qwen3-0.6B

'stem' in stem cell phrases

A neuron in layer 9 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 word 'stem' when it appears in 'stem cells', and on a few other biology words such as 'retrotransposons'. Turning it up made the model write about problems and solutions in general, which is not clearly about biology.

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

    will alter their functions accordingly. In embryonic development, stem cells differentiate into organs by means of the actions of morphogens. And even cancer

  2. #2 peak

    potential to grow into any cell or tissue in the body, and so can be used to replace cells that are damaged or diseased. Researchers believe the use of stem cells has the potential for curing Alzheimer's disease, Parkinson's disease, diabetes,

  3. #3 peak

    Edinburgh, showed that the DNA in some brain cells is different from the DNA in other cells due to a genetic phenomenon known as retrotransposons.⏎Retrotransposons are small bits of genetic information that can move around and multiply themselves in

  4. #4 peak

    new tissues in an embryo or regenerate damaged tissues in an adult. However, cancer stem cells are thought to hijack the ability of normal stem cells to express MET, and thus become the cause of cancer persistence and spread to

  5. #5 peak

    of U.S. scientists has identified phosphorylated signalling proteins in human embryonic stem cells (hESCs) studying which may help understand the mechanisms that determine whether stem cells divide or differentiate

  6. #6 peak

    at Children's Hospital Boston and the department of genetics at Harvard Medical School (HMS). "It also shows, from one lung cancer to another, the cancer stem cells are not the same."⏎Cancer stem cells are a subset of cancer cells

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 number 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 2

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

    0.432 g of Linolenic acid.⏎Nutritional facts Per 100 Grams⏎Helps in improving vision⏎Beta-carotene is very much important for improving vision. Many antioxidants can be found in palm

  2. #2 peak

    = talk_to_server(parse_xml_document=True)If you read it out loud, the code "reads"

  3. #3 peak

    atory and human feeding studies suggest that if consumed at usual rates, GR can supply as much as 42% of the EAR for Vitamin A. Beta-carotene is only converted into Vitamin A by the human body as needed<|endoftext|>Day

  4. #4 peak

    the current scenario than Roman numerals, however Roman numerals are still useful. Arabic numerals share their origin from the numbering system of

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. L8:2684 suppresses
  2. L8:2963 excites
  3. L8:600 excites
  4. L8:555 weakly suppresses
  5. L4:1158 weakly suppresses
  6. L6:16 weakly excites
  7. L5:369 weakly suppresses
  8. L4:2114 weakly suppresses
  9. L7:2892 weakly suppresses
  10. L8:2312 weakly suppresses
  11. L4:1476 weakly suppresses
  12. L7:2271 weakly excites
  13. L5:1214 weakly suppresses
  14. L7:1675 weakly excites
  15. L6:17 weakly suppresses
  16. L7:2963 weakly excites
  17. L6:1628 weakly excites
  18. L5:1438 weakly excites
  19. L0:1464 weakly suppresses
  20. L7:2648 weakly suppresses
  21. L7:2698 weakly excites
  22. L7:39 weakly excites
  23. L6:2973 weakly suppresses
  24. L7:1661 weakly suppresses

Feeds later layers

  1. L10:2678 excites
  2. L17:2487 suppresses
  3. L10:2052 suppresses
  4. L10:2186 suppresses
  5. L17:1758 suppresses
  6. L11:201 suppresses
  7. L10:1084 excites
  8. L11:1200 weakly excites words that signal persistence (still, retains, continue to)
  9. L10:1357 weakly excites
  10. L18:1848 weakly excites
  11. L20:1717 weakly excites
  12. L15:1562 weakly excites
  13. L10:277 weakly suppresses
  14. L17:1970 weakly excites
  15. L22:2861 weakly excites
  16. L11:1442 weakly suppresses
  17. L12:3058 weakly suppresses
  18. L12:759 weakly suppresses
  19. L11:1928 weakly excites
  20. L17:2383 weakly excites
  21. L18:1450 weakly suppresses
  22. L12:703 weakly excites
  23. L22:2232 weakly excites
  24. L17:1957 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 -1.17 and -1.121 values between -0.90 and -0.851 values between -0.67 and -0.621 values between -0.49 and -0.447 values between -0.44 and -0.4015 values between -0.40 and -0.3526 values between -0.35 and -0.3139 values between -0.31 and -0.2674 values between -0.26 and -0.22157 values between -0.22 and -0.17256 values between -0.17 and -0.12402 values between -0.12 and -0.08532 values between -0.08 and -0.03650 values between -0.03 and 0.01582 values between 0.01 and 0.06391 values between 0.06 and 0.10292 values between 0.10 and 0.15204 values between 0.15 and 0.19115 values between 0.19 and 0.2481 values between 0.24 and 0.2938 values between 0.29 and 0.3323 values between 0.33 and 0.3813 values between 0.38 and 0.423 values between 0.42 and 0.471 values between 0.47 and 0.511 values between 0.51 and 0.561 values between 0.56 and 0.601 values between 0.60 and 0.65offnegativepositiveoften

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. icol
  2. colspan
  3. cca
  4. 客户提供
  5. ��
  6. ]=[
  7. ander
  8. 抵
  9. ;c
  10. Qi

Suppresses

  1. .parseFloat
  2. estring
  3. .parseInt
  4. -parse
  5. later
  6. ată
  7. emouth
  8. parse
  9. eled
  10. Later