Neuron Explorer

Qwen3-0.6B

word fragments mid-word

A neuron in layer 15 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 later pieces of a word that has been split into parts, such as the ending of 'fertilisation' or 'restructuration'. Turning it up did not produce a clear theme in the model's output.

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

    .⏎There is a definite dearth of literature that deals primarily with Kurds in Syria (in Kurdish this region is sometimes referred to as Kurdistana Binxetê, meaning ‘below the line’ ). Most published works on the Kurds deal with

  2. #2 peak

    : The Chilkoot Brown Bear Project⏎I uttered the habitual salutation as I began the trek down from my isolated research cabin to the banks of Southeast Alaskaís Chilkoot River. Although I saw no evidence of recent bear activity through the

  3. #3 peak

    and key?.?.?.?gratuitously, to all who take the Library.?1 This is from the ?Prospectus? for Horace Mann?s The School Library, a series of books endorsed by the state Board of Education for

  4. #4 peak

    .<|endoftext|>Archive for the ‘transcendental geometry’ tag⏎The modern mathematician finds the space of three dimensions, in which our visible universe is containled, entirely too contracted for his

  5. #5 peak

    subjected to a long delay if a “cron job” happens to be running. Well almost…⏎Recently Dokuwiki gained an indexer to improve the performance of it’s search functionality. Dokuwiki doesn’t use a database – wiki pages are stored

  6. #6 peak

    , causing them to be dispersed widely by the water current. The eggs develop rapidly, hatching into larvae within 24 to 48 hours after fertilisation (8).<|endoftext|>Peter Wadhams’ book title is a play on words

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 **"T-1""s" in the of the of of of of of, in of of of of, the. 0.21 0.1 1. 1

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

    or as conditions dictate.⏎- Clipping the wings on a new queen is another attempt to prevent swarming. This procedure slows down swarming, but may resutt in the colony swarming anyway with a virgin queen.⏎- Reversing the

  2. #2 peak

    who finally expelled them from the district. The [[Amsterdam Mennonite Library (Bibliotheek en Archief van de Vereenigde Doopsgezinde Gemeente te Amsterdam)|Mennonite Archives of Amsterdam]] contain

  3. #3 peak

    foot of the structure at both ends. The four rusticated plinths bore allegorical sculptures in classical compositions.”⏎The Juan Arellano-designed bridge was heavilty damaged during the Liberation for

  4. #4 peak

    punishment on our black children is rooted in the violent history of American slavery. Having enslaved parents beat their own children was a prophylactic method to protect children from hasher beatings they would otherwise get from white slavers.⏎The “switch” has

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. L14:528 excites
  2. L14:1304 excites
  3. L8:423 excites
  4. L13:1903 excites
  5. L9:929 suppresses 'each' in 'each other'
  6. L13:74 excites
  7. L13:2580 suppresses
  8. L13:1155 excites
  9. L13:1552 suppresses
  10. L6:2069 suppresses
  11. L14:2121 excites
  12. L14:1265 excites
  13. L14:523 excites
  14. L13:2438 suppresses
  15. L14:311 suppresses
  16. L9:1960 suppresses
  17. L11:173 excites
  18. L14:1090 suppresses
  19. L10:1651 suppresses
  20. L14:830 suppresses
  21. L13:1439 suppresses end of document token
  22. L7:114 excites
  23. L14:1037 suppresses
  24. L8:98 excites

Feeds later layers

  1. L16:17 strongly excites
  2. L16:14 strongly excites
  3. L16:39 strongly excites
  4. L16:120 strongly excites
  5. L16:73 strongly excites
  6. L16:119 strongly excites
  7. L16:31 strongly excites
  8. L16:160 strongly excites
  9. L16:26 strongly excites
  10. L16:9 strongly excites
  11. L16:105 strongly excites
  12. L16:37 strongly excites
  13. L16:493 strongly excites
  14. L16:194 strongly excites
  15. L16:126 strongly excites
  16. L16:76 strongly excites
  17. L16:28 strongly excites
  18. L16:66 strongly excites
  19. L16:127 strongly excites
  20. L16:135 strongly excites
  21. L16:454 strongly excites
  22. L16:91 strongly excites
  23. L16:478 strongly excites
  24. L16:34 strongly 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 -2.69 and -2.531 values between -2.53 and -2.361 values between -2.36 and -2.203 values between -1.72 and -1.556 values between -1.39 and -1.233 values between -1.07 and -0.914 values between -0.91 and -0.7412 values between -0.74 and -0.5832 values between -0.58 and -0.42147 values between -0.42 and -0.26598 values between -0.26 and -0.101605 values between -0.10 and 0.07868 values between 0.07 and 0.23377 values between 0.23 and 0.39139 values between 0.39 and 0.5549 values between 0.55 and 0.7220 values between 0.72 and 0.8812 values between 0.88 and 1.044 values between 1.04 and 1.209 values between 1.20 and 1.362 values between 1.36 and 1.532 values between 1.53 and 1.691 values between 1.69 and 1.851 values between 1.85 and 2.012 values between 2.01 and 2.172 values between 2.17 and 2.342 values between 2.34 and 2.502 values between 2.50 and 2.661 values between 2.99 and 3.151 values between 3.63 and 3.80offnegativepositiveoften

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. 睢
  2. ильно
  3. sciously
  4. �
  5. ildo
  6. uet
  7. mpp
  8. rupt
  9. .Low
  10. yAxis

Suppresses

  1. van
  2. h
  3. olang
  4. eding
  5. ens
  6. ed
  7. hs
  8. JECT
  9. Scotia
  10. 县