Neuron Explorer

Qwen3-0.6B

unclear: mid-name word fragments

A neuron in layer 3 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on the first or middle piece of a longer word or name, mostly in encyclopedic history text. The pattern is loose and the writing it produces under steering 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

    , Erie, PA, was one of the few non-mountainous spots in<|endoftext|>The New Student's Reference Work/Alcinous⏎Alcinous (al-cĭn' o-us), a mythical king of the island of Scheria

  2. #2 peak

    into a dozen burial towers, the city's two major sanctuaries — the Temple of Bel (God) and the charming Temple of Baalshamin (god of fertility) — and a monumental arch that opens onto a 1,2

  3. #3 peak

    the Winyaw Sanitarium for the treatment of diseases of the lungs and throat. Modeled and conducted along the same lines as those in Germany, the sanitarium soon was receiving patients from all over the United States.⏎After Koch discovered

  4. #4 peak

    'just' a children's game.⏎"At Somerset's Punkie Night, at the end of October, children made punkies (lanterns) out of mangel-wurzels (a large kind of beet) and went knocking on

  5. #5 peak

    , who denounced King as a lawbreaker who had been manipulated by Communists. The situation looked bleak.⏎By then, Wonder had matured from a young harmonica-playing sensation to a chart-topping music genius lauded for his complex rhythms and

  6. #6 peak

    , Kalingrad Oblast, Russia), also called "the scattered church of Prussian Lithuania." The Mennonites living here had much to endure from the dragoons of the Prussian King Friedrich Wilhelm I, who finally expelled them from

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: GERMANIA, THE THEOROMACHY, 22STONIANOPOLE, FLUIDICS, THE ONLINE ### 2024, 22STONION A... 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

    ore of philadelphia who donated two chromebooks to kids at the gloucester As covid 19 cases surge across ontario hundreds of elementary students across the toronto district school board have been Children who attend preschool enter kindergarten with improved skills but

  2. #2 peak

    and traditions of their people⏎- they organized forces to win independence⏎what happened in the russo-japanese war and what was the outcome?⏎- japan crushed russia⏎- japan overran most of southeast asia during WWII, from indoch

  3. #3 peak

    the silver triumph of freedom…⏎Douglass marked down a gradual process of how his psyche driven to

  4. #4 peak

    to all of the courses in the specialization, and you earn a certificate when you complete the work.⏎for college students (except those in canada), log in to sapling learning, your online homework solution. in this course, you will learn about

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. L1:71 strongly excites
  2. L2:3 strongly suppresses
  3. L1:77 strongly excites
  4. L2:495 strongly excites
  5. L2:201 excites
  6. L2:381 excites
  7. L0:138 suppresses
  8. L1:655 suppresses
  9. L0:212 excites
  10. L0:130 excites unclear: word right after a modifier
  11. L1:1154 suppresses
  12. L2:882 excites
  13. L2:59 excites
  14. L1:762 suppresses
  15. L1:893 suppresses
  16. L2:483 suppresses
  17. L1:824 suppresses
  18. L2:969 suppresses
  19. L2:2247 suppresses
  20. L1:122 excites 'World War II' and the 'II' after 'War'
  21. L0:115 suppresses
  22. L2:74 weakly excites
  23. L2:70 weakly excites
  24. L0:1301 weakly suppresses

Feeds later layers

  1. L4:176 strongly excites
  2. L4:56 strongly suppresses
  3. L10:482 suppresses
  4. L8:637 suppresses
  5. L5:429 suppresses
  6. L7:280 suppresses
  7. L9:2624 excites
  8. L4:158 suppresses
  9. L4:1601 suppresses
  10. L9:450 suppresses
  11. L6:2656 suppresses
  12. L5:437 suppresses
  13. L5:2508 suppresses
  14. L6:505 suppresses
  15. L4:1414 suppresses
  16. L4:662 excites
  17. L4:577 suppresses
  18. L6:2959 excites
  19. L4:0 suppresses
  20. L5:1286 excites
  21. L7:272 suppresses
  22. L6:2685 suppresses
  23. L4:1740 suppresses
  24. L20:987 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 -0.77 and -0.703 values between -0.70 and -0.627 values between -0.62 and -0.5414 values between -0.54 and -0.4661 values between -0.46 and -0.39108 values between -0.39 and -0.31280 values between -0.31 and -0.23451 values between -0.23 and -0.15628 values between -0.15 and -0.081051 values between -0.08 and 0.00781 values between 0.00 and 0.08242 values between 0.08 and 0.1698 values between 0.16 and 0.2355 values between 0.23 and 0.3129 values between 0.31 and 0.3912 values between 0.39 and 0.4618 values between 0.46 and 0.5412 values between 0.54 and 0.6210 values between 0.62 and 0.705 values between 0.70 and 0.7711 values between 0.77 and 0.854 values between 0.85 and 0.936 values between 0.93 and 1.019 values between 1.01 and 1.083 values between 1.08 and 1.161 values between 1.24 and 1.323 values between 1.32 and 1.391 values between 1.39 and 1.472 values between 1.70 and 1.781 values between 1.93 and 2.011 values between 2.24 and 2.32offnegativepositiveoften

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. ronym
  2. uteur
  3. ующ
  4. itative
  5. ponsible
  6. unist
  7. reated
  8. ourced
  9. INCIDENTAL
  10. reating

Suppresses

  1. 痣
  2. tran
  3. 不舒服
  4. VILLE
  5. .port
  6. ford
  7. Isle
  8. lle
  9. XmlDocument
  10. 岛