Neuron Explorer

Qwen3-0.6B

'into' or 'to' after verbs of placing

A neuron in layer 11 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 'into' or 'to' after a verb of putting, placing or giving something somewhere, like 'placed into the Royal Navy' or 'sold into slavery'. Turning it up did not produce any clearly related writing.

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

    to the majority of the captured Meccans at the Battle of Badr?⏎They were let go⏎They were tortured⏎They were executed⏎They were sold into slavery⏎Copyright, FunTrivia.com. All Rights Reserved.⏎Legal / Conditions of

  2. #2 peak

    city was captured by the British in September 1776 and remained in their hands until 1783.⏎Four months after New York was returned to the victorious Patriots, the city was declared to be the capital of the United States

  3. #3 peak

    captured by Britain's Royal Navy during the American Revolution, and which later sank while in Royal Navy service. The captured American privateer Who's Afraid was placed into the Royal Navy and renamed HMS Mentor. The captured American privateer Independence was placed

  4. #4 peak

    and other physical responses. All the body’s combined reactions to stress are also known as ”fight or flight” responses. Our body puts all of it’s resources into what it needs to do in order to deal with the perceived stress that’s occurring

  5. #5 peak

    , a tube of fennel. He then took this flame to Earth, and gave it to men.⏎We have analysed the Prometheus myth, unpicking some of the misconception surrounding

  6. #6 peak

    subsequently became known as the Cotentin Peninsula.⏎The town was destroyed by invading Normans in 866, who later established settlements and incorporated the whole peninsula into the Duchy of Normandy in 933. On July 1

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 want to find the number of the $n$...". So, I need to know the number of... (blank) (blank) (blank) (blank)... in the word "A" in the word "B"

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

    iful eye shall be blessed; for he giveth of his bread to the poor.” – Proverbs 22:9⏎2

  2. #2 peak

    Many times, as parents (and other influences), we will need to take the reigns to teach children about these qualities. Generosity usually begins with children sharing their possessions with other peers. Not only can you learn this important quality from home,

  3. #3 peak

    interesting question that I haven’t seen anyone else ask: “If you were forced to return to your home country by the U.S. government, would you bring your children born in the U.S. with you or would you leave them in this

  4. #4 peak

    cloth and creating beautiful designs and combinations of colours.⏎India was not the only source of fine fabrics. In 1697, these silk samples were sent from the Company's factory, or base, in the Persian Gulf so that the Company

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. L9:2404 excites
  2. L6:3038 weakly excites
  3. L5:1161 weakly excites
  4. L8:1602 weakly excites
  5. L5:1746 weakly excites
  6. L6:2926 weakly excites
  7. L10:1401 weakly excites
  8. L4:1210 weakly suppresses
  9. L5:2577 weakly suppresses
  10. L7:2557 weakly suppresses
  11. L4:2718 weakly excites
  12. L7:893 weakly excites
  13. L10:1624 weakly suppresses
  14. L9:1571 weakly suppresses
  15. L4:1851 weakly suppresses
  16. L10:2657 weakly excites
  17. L6:1890 weakly excites
  18. L6:2163 weakly suppresses
  19. L6:3037 weakly suppresses
  20. L9:2147 weakly excites
  21. L1:710 weakly excites
  22. L3:1901 weakly suppresses
  23. L10:1600 weakly suppresses
  24. L6:2820 weakly suppresses

Feeds later layers

  1. L12:2459 strongly excites
  2. L16:1298 excites
  3. L16:2119 excites
  4. L12:3056 excites
  5. L12:2014 excites
  6. L15:1328 excites
  7. L12:1404 excites
  8. L16:2243 excites
  9. L12:2755 excites
  10. L13:1280 excites
  11. L14:1013 excites
  12. L12:845 excites
  13. L17:2375 excites
  14. L12:2850 excites
  15. L12:819 excites
  16. L14:2634 excites
  17. L17:2959 excites
  18. L18:1245 excites
  19. L12:2145 suppresses
  20. L17:2527 excites
  21. L22:1277 weakly excites
  22. L14:2972 weakly excites
  23. L14:1853 weakly excites
  24. L16:1233 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 -2.56 and -2.401 values between -1.93 and -1.771 values between -1.77 and -1.613 values between -1.29 and -1.137 values between -1.13 and -0.975 values between -0.97 and -0.817 values between -0.81 and -0.6527 values between -0.65 and -0.5043 values between -0.50 and -0.34206 values between -0.34 and -0.181008 values between -0.18 and -0.022034 values between -0.02 and 0.14403 values between 0.14 and 0.3095 values between 0.30 and 0.4621 values between 0.46 and 0.628 values between 0.62 and 0.787 values between 0.78 and 0.9312 values between 0.93 and 1.098 values between 1.09 and 1.252 values between 1.25 and 1.411 values between 1.57 and 1.731 values between 2.05 and 2.211 values between 2.21 and 2.363 values between 2.36 and 2.521 values between 2.52 and 2.681 values between 3.00 and 3.161 values between 3.63 and 3.79offnegativepositiveoften

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. into
  2. .ibm
  3. 臺
  4. stacks
  5. pageCount
  6. _into
  7. 台
  8. 果树
  9. })⏎⏎
  10. eil

Suppresses

  1. 知道
  2. onym
  3. .Skip
  4. unci
  5. Started
  6. Ả
  7. fois
  8. adio
  9. inal
  10. onus