Neuron Explorer

Qwen3-0.6B

document ends and record boundaries

A neuron in layer 14 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 at the seams between separate documents, such as the end-of-text marker or a line break just before a new record header. The strongest cases come from catalogue entries and text that stops abruptly. Turning it up did not produce a clear theme in the writing.

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

    O’Neil was declared a conservation area.⏎Neil burn Drive was created.⏎Canmore place was erected.<|endoftext|>ERIC Number: ED379961⏎Record Type: RIE⏎Publication Date: 19

  2. #2 peak

    ⏎- Other factors (e.g. extremes in temperatures, sunlight, tight-fitting clothing, etc.)<|endoftext|>ERIC Number: ED348648⏎Record Type: Non-Journal⏎Publication Date: 1992-Apr

  3. #3 peak

    banks, insurance companies, utility companies, and other institutions, and they pass on those costs to customers.⏎A foundation-supported initiative allows Rwandan farmers to access markets<|endoftext|>The small rufous-sided pygmy tyrant is a member of the

  4. #4 peak

    . The polyps are generally removed during a hysteroscopy because the doctor has the best possible view of them during such a procedure.⏎Some women may experience<|endoftext|>Sandrine Ceurstemont, video producer⏎How does an octopus locate

  5. #5 peak

    he doubted it would have much impact. “There’s no promise of funding,” Turner said.<|endoftext|>This article touches on cultural, linguistic, and technical issues that arise in video recording

  6. #6 peak

    about it<|endoftext|>Press Release 07-030⏎Survey Reveals Family Ties

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 the the first to have been the first to have been the themselves first and then I have been the first of themselves, so I have been the first of themselves... and so on... I have been the first of themselves,

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

    the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license is included in the page “GNU Free Documentation License”.⏎The copyright and license notices on

  2. #2 peak

    , access via your institution.⏎Buy single article⏎Instant access to the full article PDF.⏎Price includes VAT (USA)⏎Tax calculation

  3. #3 peak

    did so far:⏎Originally Posted by MrFujin⏎Enter a 5-charcter string: hgout

  4. #4 peak

    , please share it through the social media buttons.⏎When Mary arrived and saw Jesus, she fell at his feet and said, “Lord, if only you had been here, my brother would not have died.”⏎“Where have you put him?”

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. L11:623 weakly excites
  2. L2:2493 weakly excites
  3. L5:1462 weakly suppresses
  4. L8:950 weakly excites
  5. L4:1862 weakly suppresses
  6. L9:979 weakly suppresses
  7. L2:1991 weakly excites
  8. L3:2345 weakly suppresses
  9. L12:2617 weakly excites
  10. L9:81 weakly suppresses
  11. L6:1256 weakly suppresses
  12. L5:415 weakly suppresses
  13. L13:987 weakly excites
  14. L13:2057 weakly excites
  15. L10:2830 weakly suppresses
  16. L9:2290 weakly suppresses
  17. L7:743 weakly excites
  18. L10:213 weakly suppresses
  19. L6:2513 weakly suppresses
  20. L7:1631 weakly suppresses
  21. L4:1519 weakly suppresses
  22. L10:507 weakly suppresses
  23. L4:21 weakly excites
  24. L6:2082 weakly suppresses

Feeds later layers

  1. L15:1605 suppresses
  2. L19:2210 excites
  3. L16:2306 weakly excites
  4. L26:223 weakly suppresses
  5. L21:2718 weakly excites
  6. L27:1347 weakly excites
  7. L25:300 weakly suppresses
  8. L25:949 weakly excites
  9. L19:1701 weakly excites
  10. L17:2944 weakly excites
  11. L16:1310 weakly excites
  12. L15:1199 weakly suppresses
  13. L26:373 weakly suppresses
  14. L24:1578 weakly suppresses
  15. L19:1703 weakly suppresses
  16. L20:166 weakly suppresses
  17. L27:785 weakly excites
  18. L15:1567 weakly suppresses
  19. L19:2977 weakly excites
  20. L27:1442 weakly suppresses
  21. L17:230 weakly suppresses
  22. L21:1460 weakly excites
  23. L17:2316 weakly excites
  24. L15:2449 weakly suppresses line breaks mid-sentence

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.

3 values between -0.55 and -0.526 values between -0.52 and -0.493 values between -0.49 and -0.4610 values between -0.46 and -0.426 values between -0.42 and -0.3914 values between -0.39 and -0.3623 values between -0.36 and -0.3340 values between -0.33 and -0.3058 values between -0.30 and -0.2679 values between -0.26 and -0.23106 values between -0.23 and -0.20105 values between -0.20 and -0.17155 values between -0.17 and -0.14228 values between -0.14 and -0.10279 values between -0.10 and -0.07309 values between -0.07 and -0.04349 values between -0.04 and -0.01420 values between -0.01 and 0.02339 values between 0.02 and 0.06320 values between 0.06 and 0.09261 values between 0.09 and 0.12215 values between 0.12 and 0.15161 values between 0.15 and 0.19120 values between 0.19 and 0.2286 values between 0.22 and 0.2565 values between 0.25 and 0.2851 values between 0.28 and 0.3129 values between 0.31 and 0.3518 values between 0.35 and 0.3824 values between 0.38 and 0.4112 values between 0.41 and 0.446 values between 0.44 and 0.471 values between 0.47 and 0.511 values between 0.51 and 0.541 values between 0.54 and 0.573 values between 0.63 and 0.671 values between 0.67 and 0.701 values between 0.70 and 0.73offnegativepositiveoften

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. chi
  2. abwe
  3. pear
  4. pired
  5. iri
  6. Küche
  7. حتياج
  8. ซอ
  9. _SUPPLY
  10. elli

Suppresses

  1. OMET
  2. -------⏎
  3. ilent
  4. ___⏎⏎
  5. >All
  6. ==========⏎
  7. __________________⏎⏎
  8. =========⏎
  9. kneeling
  10. ---------------⏎