Neuron Explorer

Qwen3-0.6B

unclear: document ends and species names

A neuron in layer 16 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 is mixed. It fires on the invisible marker where one web page ends and the next begins, and it is pushed down by odd spellings inside scientific names such as 'Phormidium'. Turning it up gave unrelated, wandering text, so it does not show a clear theme.

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

    After Effects using precomps!⏎Check back next week to learn the difference between frame rate and shutter speed, or view more lessons at the Vimeo Video School .<|endoftext|>Some time back I presented the untold story of Kebara 2, the

  2. #2 peak

    and they wanted to do their Solar System puzzle each day as well!⏎I also highly suggest using our 100 page Solar System Printable with this unit.<|endoftext|>FORM AND

  3. #3 peak

    Test which will give us reams of data to analyze to improve those responses and preparations. These tests also help us find the weak link in the chain of complex<|endoftext|>- THE

  4. #4 peak

    increased amp-hour capacity from a parallel connection, understanding how these are different and the appropriate manner of charging your battery system is important in maximizing battery life and performance.<|endoftext|>A new set of MCQ Questions with Answer in the Computer Networking category. As

  5. #5 peak

    tool. Follow this same procedure to use them. They take measurements fixed on their grid direction, even if the tool endpoints are not perfectly aligned to the grid.<|endoftext|>Khalistan Issue: A brief Analysis⏎Author: Isha Pandey⏎

  6. #6 peak

    so happy to have you visit my site!⏎I hope that the travel guides, fun facts, and photography you find here will inspire you to explore new places!<|endoftext|>Will there be radiation when the mobile phone signal is shielded, and is it

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 question of the economic policy, but I can't find a way to solve it. But this is not a problem for me, but I can't find a way to solve it. But that's not a problem for me, but

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

    and form biofilms that cover vast areas of water. Find out what happens when other microbes hitch a ride.⏎Get more reference information on Phormidium pseudpriestleyi.⏎Phormidium pseudpriestley

  2. #2 peak

    Type INormal Type IANormal, air-entraining Type IIModerate sulfate resistance Type IIAModerate sulfate resistance, air-entraining Type II (MH) Moderate heat of hydration and moderate sulfate resistance⏎Types

  3. #3 peak

    thousand would fit on a pinhead - but they grow in groups to form hair-like structures. Find out more.⏎The first samples of Phormidium pseudpriestleyi were collected in 1907 by scientists on

  4. #4 peak

    of Missions for the Pacific, located in San Francisco (WBMP), and the Woman's Board of Missions for the Pacific Islands, located in Hawaii (WBMPI). The three mainland boards cooperated with the maleled American Board of

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. L13:2410 strongly suppresses
  2. L6:1846 excites
  3. L4:47 suppresses
  4. L6:1074 suppresses
  5. L6:1151 excites
  6. L6:1100 suppresses
  7. L13:1964 weakly excites
  8. L11:42 weakly suppresses
  9. L6:293 weakly suppresses
  10. L7:9 weakly suppresses
  11. L8:62 weakly suppresses
  12. L14:248 weakly suppresses
  13. L12:1377 weakly suppresses
  14. L3:2660 weakly excites
  15. L15:37 weakly excites
  16. L5:1127 weakly excites
  17. L11:1635 weakly excites
  18. L12:2995 weakly excites
  19. L13:1100 weakly suppresses
  20. L15:785 weakly suppresses 'of' in set phrases like 'of course'
  21. L7:1644 weakly suppresses 'to' after a noun that needs a purpose clause
  22. L1:1384 weakly suppresses
  23. L3:2458 weakly excites
  24. L6:1058 weakly suppresses

Feeds later layers

  1. L19:2985 strongly suppresses
  2. L19:541 suppresses
  3. L19:2029 suppresses
  4. L18:767 suppresses
  5. L17:259 suppresses
  6. L18:1366 suppresses
  7. L24:774 suppresses
  8. L25:1175 suppresses
  9. L18:1708 suppresses
  10. L17:1566 suppresses
  11. L22:247 suppresses
  12. L23:290 suppresses
  13. L21:732 suppresses
  14. L24:1001 suppresses
  15. L20:131 suppresses
  16. L17:264 suppresses
  17. L20:272 suppresses
  18. L20:1476 suppresses
  19. L17:1423 weakly suppresses
  20. L20:1458 weakly suppresses
  21. L18:2557 weakly suppresses
  22. L17:163 weakly suppresses
  23. L26:380 weakly suppresses
  24. L19:812 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 -2.75 and -2.452 values between -2.15 and -1.852 values between -1.85 and -1.554 values between -1.55 and -1.256 values between -0.95 and -0.6540 values between -0.65 and -0.35936 values between -0.35 and -0.052420 values between -0.05 and 0.25449 values between 0.25 and 0.5538 values between 0.55 and 0.852 values between 0.85 and 1.151 values between 2.05 and 2.352 values between 2.65 and 2.951 values between 2.95 and 3.251 values between 3.55 and 3.851 values between 6.85 and 7.151 values between 7.45 and 7.751 values between 8.95 and 9.25offnegativepositiveoften

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. { ⏎
  2. {o
  3. WD
  4. phinx
  5. غزة
  6. 扆
  7. 葬
  8. {@
  9. 坎
  10. peated

Suppresses

  1. nEnter
  2. getView
  3. 病毒感染
  4. estar
  5. ьте
  6. @register
  7. chemas
  8. CursorPosition
  9. esan
  10. 查看详情