Neuron Explorer

Qwen3-0.6B

'for' giving a reason or cause

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

The neuron fires on the word 'for' when it gives a reason, like 'killed just for fun', 'sent off for headbutting', or 'jail for a conviction'. Its negative side is weak and mixed. Turning it up did not produce text about reasons or causes.

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

    stand to see animals killed just for fun. Her sense of injustice makes her angry and through her anger she develops a

  2. #2 peak

    that took place at Wounded Knee, Mr. Camp was the only one to serve time. He spent two years in prison in Leavenworth, Kan., for assaulting a postal inspector, a charge that Camp-Horinek disputes.⏎In

  3. #3 peak

    May to July 2006, and towards the end of that period France lost the World Cup final after Zinedine Zidane was sent off for headbutting an Italian player. Despite the national defeat,<|endoftext|>By Connie Vit

  4. #4 peak

    The Magic Finger.⏎I can't stand hunting. I just can't stand it. It doesn't seem right to me that men and boys should kill animals just for the fun they get out of it.⏎The Girl 1⏎The Magic Finger

  5. #5 peak

    earlier admission that the Church made a⏎grievous error when it persecuted Galileo for arguing that the⏎Earth revolved around the sun. The article further described the

  6. #6 peak

    the last 25 years or so."⏎Powell interviewed Tupac a number of times for Vibe magazine, including while the rapper was serving time in jail for a sexual abuse conviction. He joined NPR's Renee Montagne to discuss Tup

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 am the teacher and I am tired of this. I am the teacher and I am tired of this. I am the teacher and I am tired of this. I am the teacher and I am tired of this. I am the

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

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

  2. #2 peak

    everyday terms, it's what is needed to move things, heat them... More into the light bulb⏎- The excess energy returns to the breaker panel through a neutral wire, completing a circuit.⏎This process continues until you turn off the light bulb

  3. #3 peak

    Argentina.⏎Historically, lithium has been mined either in open pits or underground or extracted from brine using large, plastic-lined solar evaporation ponds. Producing lithium from evaporation ponds can take months and has drawn criticism from some environmental groups for

  4. #4 peak

    and then its right child.⏎T / \ I S / \ D E<|endoftext|>Original URL: http://www.theregister.co.uk/2006/08/21/mars_geysers/⏎Martian pole freck

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. L5:1090 excites
  2. L6:2553 excites 'for' after praise words like 'known for'
  3. L6:1943 weakly excites
  4. L3:733 weakly excites
  5. L8:2745 weakly suppresses
  6. L6:2143 weakly excites
  7. L6:1671 weakly excites
  8. L6:2941 weakly suppresses
  9. L4:1243 weakly suppresses
  10. L5:60 weakly excites
  11. L5:1285 weakly suppresses
  12. L0:1543 weakly excites
  13. L3:2019 weakly suppresses
  14. L3:381 weakly excites 'for' in research institute names
  15. L7:3049 weakly suppresses
  16. L5:216 weakly suppresses
  17. L4:2502 weakly suppresses
  18. L3:2322 weakly suppresses
  19. L1:784 weakly excites
  20. L5:110 weakly suppresses
  21. L3:163 weakly suppresses
  22. L0:1356 weakly excites
  23. L4:1395 weakly excites
  24. L5:1213 weakly suppresses

Feeds later layers

  1. L10:1855 excites
  2. L13:2509 excites
  3. L24:2492 weakly excites
  4. L23:331 weakly excites
  5. L22:361 weakly excites
  6. L10:1054 weakly excites 'from' after movement or transfer verbs
  7. L17:2931 weakly excites
  8. L10:2075 weakly excites
  9. L15:1642 weakly excites
  10. L14:965 weakly excites
  11. L12:2694 weakly excites unclear: 'purpose' and words near it
  12. L18:2859 weakly excites 'for' after an award or honour
  13. L20:56 weakly excites
  14. L14:2288 weakly suppresses
  15. L18:2238 weakly suppresses
  16. L15:1273 weakly excites
  17. L24:2885 weakly suppresses
  18. L11:821 weakly excites
  19. L15:765 weakly suppresses
  20. L10:2311 weakly excites indefinite article before a noun
  21. L15:1933 weakly excites
  22. L10:871 weakly excites
  23. L26:1081 weakly excites
  24. L21:2337 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 -1.71 and -1.512 values between -0.94 and -0.758 values between -0.75 and -0.5691 values between -0.56 and -0.37750 values between -0.37 and -0.181924 values between -0.18 and 0.02945 values between 0.02 and 0.21130 values between 0.21 and 0.4027 values between 0.40 and 0.598 values between 0.59 and 0.784 values between 0.78 and 0.973 values between 0.97 and 1.162 values between 1.16 and 1.363 values between 1.36 and 1.552 values between 1.55 and 1.741 values between 1.74 and 1.932 values between 2.31 and 2.501 values between 2.50 and 2.691 values between 2.89 and 3.081 values between 3.84 and 4.031 values between 4.03 and 4.221 values between 5.75 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. 引起的
  2. asp
  3. avigate
  4. 敌
  5. Dirty
  6. avig
  7. _));⏎
  8. due
  9. PostBack
  10. legitimate

Suppresses

  1. ورة
  2. imators
  3. acements
  4. 不多
  5. invers
  6. ワーク
  7. .icons
  8. icus
  9. essor
  10. anguages