Neuron Explorer

Qwen3-0.6B

ordinal numbers like 16th or 1st

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

This neuron lights up on the ending of ordinal numbers such as '16th century' or '38th president', the part that comes after the digits. Turning it up did not produce a clear topic in the model's writing; the output mostly drifted into unrelated word problems and loops.

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

    most of his life to learning and implementing the tsujigahana, a 16th-century technique that produces the most obscure and jaw-dropping masterpieces. He used

  2. #2 peak

    Yoshitoshi. It depicts a difficult battle from the Taiheiki (Chronicle of Great Peace), a Japanese historical epic written in the late 14th century and covering the period from 1319 to 136

  3. #3 peak

    is traditionally focused on repentance. Speaking biblically, to repent means to make a change in our attitudes, words, and lifestyles. As 16th century reformer Martin Luther taught, the Christian life in its totality is a

  4. #4 peak

    um), held between 1545 and 1563 in Trent (or Trento), now in northern Italy, was the 19th ecumenical council of the Catholic Church. Prompted by the Protestant Re

  5. #5 peak

    oke from the ancient myths in mind: That people actually have been named Loke or Lokke:⏎Among the Norsemen in Northumberland in the 12th century, there was a man called Locchi.⏎In Scandia, Lok

  6. #6 peak

    seven roles in his job, from being the Commander-In-Chief to the Guardian of the Economy.⏎Brief biography of Gerald R. Ford, the 38th president of the United States of America.⏎Timeline of the life of former U.S

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: the power of a plant that is not in a container, but just as in the container. So, what do you think that is the most important thing? I have to keep this in mind for me, I think, it's a

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

    Rivne region, about 200 kilometres from the site of the Chernobyl explosion in 1986.⏎They found levels of radioactive caesium in milk above Ukraine’s safe limit for adults of 100 Becquer

  2. #2 peak

    more than 12 times that for children.⏎“More than 30 years after the Chernobyl disaster, people are still routinely exposed to radioactive caesium when consuming locally produced staple foods, including milk, in Chernobyl-affected

  3. #3 peak

    children are the main consumers of that milk,” said Labunska.⏎“Though the level of soil contamination in the studied areas is not extremely high, radioactive caesium continues to accumulate in milk and other foods, such that the residents of these villages

  4. #4 peak

    ic, and are considered to be a primary contributor of arsenic in the diet. Shellfish may also contain methylmercury and certain radioactive isotopes of cesium due to leakage from the Fukushima power plant, depending on origin. See: Ar

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. L4:2364 strongly suppresses
  2. L3:2400 excites
  3. L1:2787 suppresses
  4. L6:411 suppresses
  5. L8:785 suppresses
  6. L3:2600 suppresses
  7. L6:2402 weakly excites
  8. L3:2943 weakly suppresses
  9. L4:867 weakly suppresses
  10. L3:2780 weakly excites
  11. L5:2936 weakly excites
  12. L7:1810 weakly excites
  13. L2:1221 weakly excites
  14. L6:1119 weakly suppresses
  15. L2:2107 weakly excites
  16. L7:2723 weakly excites
  17. L2:911 weakly suppresses
  18. L3:2720 weakly excites
  19. L5:2455 weakly suppresses
  20. L1:420 weakly excites
  21. L6:621 weakly suppresses
  22. L0:1111 weakly excites
  23. L1:572 weakly excites
  24. L6:2296 weakly suppresses

Feeds later layers

  1. L18:2109 excites
  2. L12:637 excites
  3. L14:1601 excites
  4. L17:1594 excites
  5. L22:2368 excites
  6. L25:2232 excites
  7. L10:1744 excites
  8. L23:465 excites
  9. L21:939 excites
  10. L13:1006 excites
  11. L25:2583 excites
  12. L14:2281 excites
  13. L11:1383 weakly suppresses
  14. L11:233 weakly excites
  15. L16:2776 weakly excites
  16. L10:1623 weakly excites
  17. L17:1628 weakly excites
  18. L17:1858 weakly excites
  19. L18:1662 weakly excites
  20. L11:2781 weakly excites
  21. L25:2100 weakly excites
  22. L20:2953 weakly excites
  23. L25:2234 weakly excites
  24. L25:1220 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 -3.43 and -3.201 values between -1.43 and -1.211 values between -1.21 and -0.991 values between -0.99 and -0.768 values between -0.76 and -0.5494 values between -0.54 and -0.321070 values between -0.32 and -0.102282 values between -0.10 and 0.12411 values between 0.12 and 0.3522 values between 0.35 and 0.574 values between 0.57 and 0.794 values between 0.79 and 1.013 values between 1.23 and 1.461 values between 2.12 and 2.341 values between 2.34 and 2.571 values between 2.57 and 2.791 values between 2.79 and 3.011 values between 3.01 and 3.231 values between 5.23 and 5.45offnegativepositiveoften

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. eenth
  2. 年第
  3. -first
  4. attribution
  5. 第一位
  6. 髻
  7. ㄡ
  8. .getFirst
  9. 第三次
  10. teenth

Suppresses

  1. )?$
  2. 耗
  3. ?"⏎⏎
  4. "?⏎⏎
  5. ibel
  6. 鲀
  7. lệ
  8. elsen
  9. ?"⏎
  10. reste