Neuron Explorer

Qwen3-0.6B

blank token before a number or after a count

A neuron in layer 6 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 reacts to an empty spot in the text, often in web page footers like "3 Views [blank] 20 Downloads" or between two numbers. It is hard to say more than that. Turning it up or down made the model's writing shift topic without a clear pattern.

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

    3: 25. doi:10.1023/A:1005462530658⏎- 896 Downloads⏎Using recently available data from China's Disease Surveillance Points system

  2. #2 peak

    well as improve their athletic skills. They work cooperatively playing soccer, badmitton, frisbee, bean bag bowling and other games.⏎17 Views 20 Downloads<|endoftext|>In Key Stage 1 the children will:⏎The five enquiry

  3. #3 peak

    explore the Orange Revolution. In this Orange Revolution lesson, students examine the climate of the Ukrainian protest and investigate the role of Canadian government in the protest.⏎3 Views 2 Downloads<|endoftext|>The environment contains those factors ‘surrounding the organization’.

  4. #4 peak

    paper, with margins; letterpress text on verso; platemark: 422 x 531 mm; total: 518 x 638 mm; watermark: Atlas holding the celestial spheres; slightly frayed

  5. #5 peak

    extraordinary 33 years in average life span over the past century. A baby born today can expect to live an average of 78 years, compared with 45 years back in 1900. But those gains have not

  6. #6 peak

    verb tense review worksheet, students are given sentences and must find the mistake. Students correct the mistakes when they find them using their knowledge of verb tense.⏎3 Views 28 Downloads<|endoftext|>Blank Electoral College Map 201

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 first thing to understand is that the first thing to understand is that the first thing to understand is to understand that the first thing to understand is to understand that the first thing to understand is to understand that the first thing to understand is to

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

    "documentation style"?⏎A documentation style is a standard, agreed-on method for creating citations. It provides formats for citing differnet information resources within the body of the paper, for listing different information types on the Works Cited page, and even

  2. #2 peak

    fascinating historical perspectives, as well as covering the role of the Irish language in Northern Ireland's more recent past. But the main body of the book is based on Malcolm's extensive and detailed research into the attitudes of young Protest

  3. #3 peak

    345 Can we simplify? No, that's not possible.⏎The answer is 315 + 35 = 345⏎Example 2: 214

  4. #4 peak

    of eqns.⏎%for any input values t, y⏎ydot = -10*y %compute rhs⏎The function needs to be tested. To do this, issue the command⏎q = rhs(0.2,

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. L5:53 weakly excites
  2. L3:1865 weakly suppresses
  3. L5:2407 weakly excites
  4. L2:70 weakly excites
  5. L4:53 weakly excites blank slot before a number
  6. L4:72 weakly suppresses
  7. L0:919 weakly excites
  8. L1:1674 weakly excites
  9. L0:21 weakly suppresses last digit of years like 1970
  10. L5:1195 weakly excites
  11. L5:46 weakly excites
  12. L0:1938 weakly suppresses
  13. L5:50 weakly excites
  14. L1:260 weakly excites
  15. L5:1939 weakly suppresses
  16. L3:2140 weakly suppresses
  17. L5:1264 weakly excites
  18. L3:230 weakly suppresses
  19. L2:155 weakly suppresses
  20. L5:322 weakly suppresses
  21. L3:1661 weakly suppresses
  22. L5:1761 weakly excites
  23. L5:285 weakly excites
  24. L5:932 weakly excites

Feeds later layers

  1. L7:2 strongly suppresses
  2. L10:17 strongly suppresses
  3. L7:24 strongly excites
  4. L11:43 strongly suppresses
  5. L9:8 strongly excites
  6. L11:41 strongly excites
  7. L11:518 excites
  8. L13:41 excites the word 'the' after a lead-in phrase
  9. L14:1962 suppresses
  10. L7:13 suppresses 'the' after 'member of' or 'one of'
  11. L12:130 excites
  12. L9:87 excites
  13. L14:1267 excites
  14. L10:42 suppresses
  15. L14:2272 excites
  16. L8:98 suppresses
  17. L10:146 weakly suppresses
  18. L7:184 weakly suppresses
  19. L18:505 weakly suppresses
  20. L16:1335 weakly suppresses
  21. L12:267 weakly suppresses
  22. L15:764 weakly suppresses
  23. L21:359 weakly excites
  24. L15:1815 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 -2.74 and -2.611 values between -2.61 and -2.491 values between -1.99 and -1.861 values between -1.86 and -1.732 values between -1.61 and -1.481 values between -1.48 and -1.361 values between -1.36 and -1.232 values between -1.11 and -0.983 values between -0.98 and -0.863 values between -0.86 and -0.732 values between -0.73 and -0.608 values between -0.60 and -0.4844 values between -0.48 and -0.35155 values between -0.35 and -0.23516 values between -0.23 and -0.101330 values between -0.10 and 0.021161 values between 0.02 and 0.15498 values between 0.15 and 0.28121 values between 0.28 and 0.4036 values between 0.40 and 0.535 values between 0.53 and 0.654 values between 0.65 and 0.784 values between 0.78 and 0.901 values between 0.90 and 1.033 values between 1.03 and 1.152 values between 1.15 and 1.281 values between 1.40 and 1.531 values between 2.16 and 2.28offnegativepositiveoften

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. noon
  2. ateurs
  3. жа
  4. usan
  5. urar
  6. .AddInParameter
  7. haul
  8. ateur
  9. 长征
  10. Ữ

Suppresses

  1. ip
  2. ertz
  3. ip
  4. butt
  5. cip
  6. 种
  7. bonded
  8. posit
  9. orneys
  10. vip