Neuron Explorer

Qwen3-0.6B

the digit 9 at a fragment start

A neuron in layer 0 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 reacts to the digit 9 when it is the first character of a piece of text, often the start of a year or number. Turning it up produced no obvious theme in the output.

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

    9 testing relying on micro/nano technology.⏎- Micro/nano technologies for drug analysis

  2. #2 peak

    932. The hotel was used as a seminary for the Society of Jesus

  3. #3 peak

    9 January 1875 to Lizzie Mark and her husband, Henry Benjamin

  4. #4 peak

    92 she passed the university junior scholarship examination.⏎In 1893 Eth

  5. #5 peak

    9-0.965). Children were more vulnerable to rotavirus (p

  6. #6 peak

    990 to 2002. It was the first effigy to

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 have been thinking about lately, and I have been thinking about lately, and I have been thinking about lately. Maybe you have been thinking about lately, and I have been thinking about lately. So, the problem is that I have

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

    most developed economies, is home to The Beatles, the birthplace of rock and pop music. England is also a site of numerous ancient Stonehenge, Baths of Stonehenge and thousands of years-old castles and churches. The Welsh valleys

  2. #2 peak

    with Washington Irving and Samuel Taylor Coleridge.⏎Returning to England, he was very influenced by romantic artists like

  3. #3 peak

    toplasm of these cells majorly is covered by the granules, excluding the bottom. The basal portion comprise of nucleus, mitochondria, endoplasmic reticulum and the other organelles. Microvilli are projected by

  4. #4 peak

    his personal diary. He was not the only US President to grow hemp however, Thomas Jefferson, James Madison, James Monroe, Andrew Jackson, Zachary Taylor, and Franklin Pierce were all reported to have sown the hemp seed!

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

None: this is the first layer.

Feeds later layers

  1. L1:1619 excites the very first token '0'
  2. L1:110 excites
  3. L1:1662 excites
  4. L1:81 suppresses
  5. L1:460 excites
  6. L21:1884 weakly suppresses
  7. L1:1337 weakly excites
  8. L2:60 weakly excites
  9. L1:218 weakly excites
  10. L1:437 weakly excites the digit 9 at a fragment start
  11. L4:278 weakly excites
  12. L26:648 weakly excites
  13. L1:1321 weakly excites
  14. L22:1654 weakly excites
  15. L1:2147 weakly excites
  16. L1:1262 weakly excites
  17. L17:2834 weakly excites
  18. L1:1431 weakly excites
  19. L26:21 weakly excites
  20. L17:2268 weakly excites
  21. L10:2459 weakly excites
  22. L2:128 weakly excites
  23. L17:1433 weakly excites
  24. L19:706 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.

2 values between -0.44 and -0.353 values between -0.35 and -0.2743 values between -0.27 and -0.18251 values between -0.18 and -0.101172 values between -0.10 and -0.011968 values between -0.01 and 0.08438 values between 0.08 and 0.1624 values between 0.16 and 0.254 values between 0.25 and 0.331 values between 1.36 and 1.452 values between 2.90 and 2.99offnegativepositiveoften

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. variable
  2. moments
  3. outcomes
  4. solid
  5. long
  6. best
  7. moment
  8. )[-
  9. aff
  10. (L

Suppresses

  1. leases
  2. Submission
  3. äß
  4. _submission
  5. arkan
  6. utron
  7. mdb
  8. lder
  9. iedade
  10. 入口