Neuron Explorer

Qwen3-0.6B

second digit of years 20xx

A neuron in layer 4 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 '0' in the second half of a year written as 20 followed by a digit, like the '0' in 2010. The model did not show a clear theme when this neuron was turned up.

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

    analysis of algoritms⏎robert.kern at gmail.com⏎Fri Sep 10 00:10:22 CEST 2010⏎On 9/9/10 4:39

  2. #2 peak

    well as 551 Alpine lakes between <|endoftext|>Gonorrhoea - the drugs don't work⏎Published: 23rd Dec 2011 08:41:37⏎The prospect of unt

  3. #3 peak

    about this plant, please leave a comment. Comments have to be approved before they are shown here.<|endoftext|>Save water with wind power⏎21 March 2012⏎The power sector is the one of the world's biggest consumers of

  4. #4 peak

    I have some simple code example?⏎## Random fruit response # Lobster March 3 2012 import random verb = ['erratic', 'unbelievable', '

  5. #5 peak

    are at right angles<|endoftext|>The latest news from academia, regulators⏎research labs and other things of interest⏎Posted: September 29, 2006⏎Magnetic nanobridge brings two worlds together⏎(Nanower

  6. #6 peak

    problems we face in adulthood? (read more here)<|endoftext|>Задание от гостя:Размещено 13.09.2016 19:19

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: you have to be... But I have to be... But I have to be... But I have to be... But I have to be... But I have to be... But I have to be... But I have to be... But

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

    to the brain caused by an excess in salt intake.⏎Dr. Costantino Iadecola, of the Feil Family Brain and Mind Research Institute at Weill Cornell Medicine in New York, led the study which involved giving mice between eight and

  2. #2 peak

    schools reopening, relaxed pandemic precautions, and surging Delta variant infections.⏎Patients who receive their primary care internal medicine or pediatric care at Weill Cornell Medicine can also receive the vaccine, along with their flu vaccine at an upcoming

  3. #3 peak

    wellbeing of its constituents is also of interest, especially to some of the oldest theorists in this school, the Transcendentalists.⏎Thoreau, Whitman, and Emerson have served as inspiration for generations of non-conformists, such

  4. #4 peak

    one of the first American periodicals to recognize emerging American literature and culture, separate from European content.⏎The magazine also contained rare en<|endoftext|>F. A. Hayek, “Planning, Science and Freedom”, Nature, Nov. 15,

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. L2:92 weakly excites
  2. L0:1529 weakly excites
  3. L3:512 weakly suppresses
  4. L2:75 weakly suppresses
  5. L2:752 weakly excites
  6. L2:3068 weakly suppresses
  7. L3:272 weakly excites
  8. L1:1960 weakly excites
  9. L3:1911 weakly excites
  10. L1:2497 weakly suppresses
  11. L3:456 weakly suppresses
  12. L3:282 weakly suppresses
  13. L2:111 weakly suppresses
  14. L0:1099 weakly suppresses
  15. L2:896 weakly suppresses
  16. L1:1460 weakly excites
  17. L2:503 weakly excites
  18. L3:50 weakly excites
  19. L2:1141 weakly excites
  20. L3:244 weakly excites
  21. L3:2954 weakly suppresses
  22. L1:1357 weakly excites
  23. L2:2283 weakly suppresses
  24. L2:1321 weakly suppresses

Feeds later layers

  1. L8:457 excites year digits in copyright and dates
  2. L18:1976 excites
  3. L12:1264 excites gap before a year in dates
  4. L9:929 weakly excites 'each' in 'each other'
  5. L7:1835 weakly suppresses
  6. L20:548 weakly excites
  7. L17:1050 weakly excites
  8. L6:2643 weakly excites
  9. L11:2368 weakly suppresses
  10. L7:1953 weakly suppresses
  11. L13:1536 weakly excites
  12. L7:676 weakly suppresses
  13. L14:197 weakly suppresses
  14. L6:215 weakly excites
  15. L10:1525 weakly excites
  16. L5:2502 weakly suppresses
  17. L5:572 weakly suppresses
  18. L5:10 weakly excites
  19. L20:3052 weakly suppresses
  20. L15:1940 weakly suppresses
  21. L23:1565 weakly suppresses
  22. L23:2100 weakly suppresses
  23. L26:1459 weakly excites
  24. L19:1273 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 -5.12 and -4.847 values between -0.97 and -0.6925 values between -0.69 and -0.41564 values between -0.41 and -0.142865 values between -0.14 and 0.14418 values between 0.14 and 0.4210 values between 0.42 and 0.691 values between 0.69 and 0.972 values between 0.97 and 1.251 values between 1.25 and 1.521 values between 1.52 and 1.801 values between 1.80 and 2.081 values between 2.35 and 2.631 values between 2.63 and 2.911 values between 2.91 and 3.182 values between 3.18 and 3.461 values between 3.46 and 3.742 values between 4.29 and 4.571 values between 4.57 and 4.841 values between 5.12 and 5.402 values between 5.67 and 5.95offnegativepositiveoften

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. iaux
  2. oons
  3. igue
  4. omen
  5. ibur
  6. iture
  7. IOUS
  8. igans
  9. oreal
  10. ṇ

Suppresses

  1. .ConnectionString
  2. 技术服务
  3. maint
  4. 上课
  5. 对此
  6. prostitution
  7. strtol
  8. QT
  9. отношении
  10. categorical