Neuron Explorer

Qwen3-0.6B

'of' after made up, composed, consists

A neuron in layer 8 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 'of' (or 'from') right after a verb that describes what something is made of, such as 'made up of', 'composed of', 'consists of' or 'comprised of'. It seems to expect a list of parts next. Turning it up made the text ramble without a clear link to materials.

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

    old on Friday April 5th in 1974. On what day⏎of the week was she born?⏎A tetromino is made up of four squares joined edge to edge. Can⏎this tetrom<|endoftext|>Congress, C

  2. #2 peak

    an oil reserve the size of England⏎The tar sands are a source of oil buried below the Boreal Forest of northern Alberta. The tar sands are made up mostly of sand. Only 10-12

  3. #3 peak

    illnesses approximated towards itself another Vedic idea of evident practical value – that everything in the universe is composed primarily of agni (fire – the heating principle) and soma (moon – the cooling

  4. #4 peak

    one was adopted by the tribal elders. It consists of seven stripes, the central being 1/6 the size of the other stripes. The hoist consists of a triangle edged in black with an semi-circle of red over black, separated by

  5. #5 peak

    was able to maintain itself near sea level by growing upwards. That reef is now over 516 feet (160 m) thick and comprised of mostly post-Miocene limestone with a layer of upper Miocene (Tertiary g

  6. #6 peak

    not dehydrate during the roasting process, has a very small shrinkage, and has relatively stable physical and chemical properties.⏎The slurry used is mainly composed of various components of water suspension colloid. When making embryos, the slurry can

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 90% chance of success on my first attempt at the exam for the exam, and 90% on the second attempt. I have a 90% chance of success on the second attempt. What is

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

    devices(2427) Terms.⏎The official website of the Tokyo 2020 - Summer Olympic Games - 3x3 event 2021 - FIBA.basketball⏎Here to

  2. #2 peak

    functions is vital. Some of the essential facts you will need to note includes:⏎- Every function should have a unique name⏎- A function name must be followed by parentheses⏎- Functions need to equipped with return type⏎- A function body has

  3. #3 peak

    launched by Expeditions, an independent network of scholars in the human sciences. For 11 years, we have been holding in Gozo (Malta) a summer school for anthropologists and social scientists, focused on the practice of fieldwork

  4. #4 peak

    letters: A D E L M N O P S.⏎No direct anagrams for endoplasm found in this word list.⏎Words formed by adding one letter before or after endoplasm (in bold), or to adelmnops in any

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. L4:2702 weakly excites
  2. L4:2991 weakly excites
  3. L7:2688 weakly suppresses
  4. L6:13 weakly excites
  5. L2:2825 weakly excites
  6. L1:2054 weakly excites
  7. L7:500 weakly excites
  8. L6:778 weakly excites
  9. L7:2456 weakly suppresses
  10. L7:2590 weakly excites
  11. L7:2474 weakly suppresses
  12. L1:2079 weakly excites
  13. L6:2836 weakly excites
  14. L6:12 weakly excites
  15. L2:1874 weakly suppresses
  16. L3:2379 weakly excites
  17. L0:1702 weakly suppresses
  18. L6:1139 weakly suppresses
  19. L7:196 weakly suppresses
  20. L6:2377 weakly excites
  21. L4:2111 weakly excites
  22. L5:812 weakly suppresses
  23. L5:1671 weakly excites
  24. L5:470 weakly excites

Feeds later layers

  1. L9:1595 excites
  2. L12:1544 weakly excites
  3. L10:1539 weakly excites
  4. L9:796 weakly excites
  5. L17:2888 weakly excites
  6. L10:2120 weakly suppresses
  7. L16:2068 weakly excites
  8. L14:2374 weakly excites
  9. L10:1406 weakly excites
  10. L24:2111 weakly excites
  11. L12:1531 weakly suppresses
  12. L15:1836 weakly excites
  13. L12:1636 weakly excites
  14. L23:638 weakly excites
  15. L11:312 weakly excites
  16. L9:2132 weakly excites
  17. L11:600 weakly suppresses
  18. L18:2387 weakly suppresses
  19. L22:2710 weakly excites
  20. L12:1859 weakly excites
  21. L24:1048 weakly suppresses
  22. L9:881 weakly excites
  23. L10:2300 weakly excites
  24. L9:2028 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 -1.92 and -1.672 values between -1.43 and -1.184 values between -1.18 and -0.933 values between -0.93 and -0.6923 values between -0.69 and -0.44304 values between -0.44 and -0.192650 values between -0.19 and 0.05831 values between 0.05 and 0.3059 values between 0.30 and 0.5513 values between 0.55 and 0.795 values between 0.79 and 1.045 values between 1.04 and 1.281 values between 1.28 and 1.531 values between 1.78 and 2.021 values between 2.27 and 2.521 values between 3.01 and 3.261 values between 3.50 and 3.751 values between 3.75 and 4.001 values between 6.21 and 6.461 values between 7.69 and 7.94offnegativepositiveoften

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. comprised
  2. volunteers
  3. consisted
  4. 组成
  5. getContent
  6. naires
  7. composed
  8. entimes
  9. acquainted
  10. contained

Suppresses

  1. versions
  2. .Put
  3. 方向
  4. Alternate
  5. redirect
  6. ender
  7. american
  8. Directions
  9. offs
  10. secure