Neuron Explorer

Qwen3-0.6B

unclear: quantities and measured changes

A neuron in layer 13 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 reacts to words in the middle of number-based phrases, like a fraction, a rate of change, or a count of things. It is a loose pattern, and turning it up gave rambling text that had nothing to do with numbers.

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

    not until June 15, 1956 that 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

  2. #2 peak

    4% of strokes worldwide, the authors write. They note that the risk of heart attack and stroke drops by about half for every 20 mm/Hg reduction in systolic blood pressure (the top number of your blood pressure

  3. #3 peak

    lic acids. The rate at which this occurs depends on a number of factors. Temperature is perhaps the most critical one, since the rate of oxidation doubles for every rise of

  4. #4 peak

    than merely memorising them, the researchers introduced words the birds had never seen before.⏎The pigeons correctly identified the new words as words at a rate significantly above chance.⏎According to the study’s first author, Dr Damian Scarf of the University

  5. #5 peak

    | Permalink⏎"crochet school - Lesson 9", "crochet school", "learn to crochet", "treble crochet", crochet⏎| Digg This⏎| Save to del.icio.us<|endoftext|>Back in the early 2

  6. #6 peak

    , or they will invest elsewhere.⏎The following is the WACC calculation formula:⏎WACC = E/V × Re + D/V × Rd × (1 - Tc)⏎Re = cost of equity⏎Rd = cost of debt⏎E

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: 10000000000000000000000000000000000000000000000

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

    2. A company is offering to pay you one penny on your first day at work and each day after that your salary would triple. Assume you work five days a week.⏎a. Fill out the tables representing your daily salary for your first two

  2. #2 peak

    and girls. From six years onwards, you detect small differences<|endoftext|>An average size tree produces enough oxygen in one year to keep a family of four breathing.⏎Three trees planted in the right place around buildings can cut air-conditioning costs up to

  3. #3 peak

    to activate it.⏎Can vinegar and baking soda be mixed?⏎Here are some recipes to try. Freshen your sink by mixing one part of baking soda with two parts of vinegar. This mixture unlocks an eff

  4. #4 peak

    Yuan replaced the Fǎbì in 1948 at a rate of 1 Gold Yuan to 3 million Yuan Fǎbì. That same year, the Yuan Renminbi

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. L11:95 excites
  2. L12:1686 excites
  3. L11:726 suppresses
  4. L12:2169 suppresses
  5. L11:1036 suppresses
  6. L11:805 suppresses
  7. L12:1141 excites
  8. L12:983 suppresses unclear: mid-clause words in expository prose
  9. L11:2493 suppresses
  10. L11:2502 excites
  11. L12:2937 weakly suppresses
  12. L8:1827 weakly excites
  13. L10:1252 weakly suppresses
  14. L11:1775 weakly excites
  15. L11:2812 weakly excites
  16. L11:1889 weakly suppresses
  17. L12:1350 weakly excites
  18. L11:374 weakly excites
  19. L12:2085 weakly suppresses
  20. L12:1446 weakly suppresses
  21. L10:2487 weakly excites
  22. L10:114 weakly excites
  23. L12:390 weakly suppresses
  24. L11:2534 weakly suppresses

Feeds later layers

  1. L14:2850 suppresses
  2. L14:258 excites
  3. L16:1154 suppresses
  4. L17:1804 suppresses
  5. L25:646 suppresses measurement words before a value
  6. L21:1243 excites
  7. L18:2465 excites
  8. L22:2844 weakly suppresses
  9. L14:2303 weakly excites
  10. L16:2643 weakly excites
  11. L14:3003 weakly excites
  12. L15:2789 weakly excites
  13. L15:1423 weakly suppresses
  14. L14:365 weakly excites
  15. L14:1706 weakly suppresses
  16. L18:2584 weakly excites
  17. L14:1507 weakly excites
  18. L16:1491 weakly suppresses
  19. L17:481 weakly excites
  20. L16:2402 weakly excites
  21. L16:2998 weakly suppresses
  22. L14:914 weakly excites
  23. L16:2106 weakly excites
  24. L14:2832 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.

1 values between -7.16 and -6.931 values between -5.06 and -4.821 values between -4.36 and -4.121 values between -2.72 and -2.481 values between -2.25 and -2.023 values between -2.02 and -1.782 values between -1.78 and -1.551 values between -1.55 and -1.3210 values between -1.32 and -1.087 values between -1.08 and -0.8529 values between -0.85 and -0.6126 values between -0.61 and -0.38215 values between -0.38 and -0.152488 values between -0.15 and 0.09936 values between 0.09 and 0.32132 values between 0.32 and 0.5625 values between 0.56 and 0.7914 values between 0.79 and 1.026 values between 1.02 and 1.266 values between 1.26 and 1.492 values between 1.49 and 1.721 values between 1.96 and 2.19offnegativepositiveoften

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. esome
  2. IDA
  3. newList
  4. \a
  5. ows
  6. adow
  7. fixed
  8. TINGS
  9. azing
  10. ANGED

Suppresses

  1. veces
  2. оличество
  3. RefCount
  4. salt
  5. Promotion
  6. sem
  7. inness
  8. reviews
  9. يون
  10. 盐