Neuron Explorer

Qwen3-0.6B

the word 'exception' or 'exceptions'

A neuron in layer 3 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 lights up on the word 'exception' (or 'exceptions') when it comes right after words like 'the only', 'is no', or 'an'. When it was turned up, the written text drifted into unrelated topics and got repetitive, so the steering shows no clear theme.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    ,” you do it. I’m usually not a fan of historical fiction, particularly novels set during wars, but Joseph Bruchac’s Code Talker is an exception, a book I’m so happy to have been recommended.⏎During WWII, the

  2. #2 peak

    article Oils, Fats and Health).Varieties⏎General: Oils and fats are pretty uniform at 120 calories per tablespoon. The only exceptions are semi-solid products sold whipped with a lot of air - fewer calories by the

  3. #3 peak

    . When an athlete breaks the law, the league tries to avoid making a statement with their response. Michael Vick, Donald Sterling, and post-video Ray Rice are exceptions. Leniency—what we saw in the initial response to Rice’s behavior

  4. #4 peak

    s. Words ending in -quy don’t follow this rule (as in soliloquies).⏎Add -s to most nouns ending in -f.⏎Exceptions: In some cases, change the -f to -v and add -es

  5. #5 peak

    953, before there was a reasonably complete edition of his writings. Sure, there were plenty of Lincoln biographies before those dates, but with only one exception (an authorized multi-volume door-stopper by his former White House aides), there

  6. #6 peak

    indication of the part of the wall upon which they labored, such places being on that side of the city nearest their place of abode; the only apparent exception being, perhaps, where they repaired more than their

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 working on this for a while, and I have been thinking about it in my mind, so I think that this is a good experience. I have been thinking about this for a while, and I think that this is 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

    don’t like to eat food contaminated by pesticides,” said EWG president Ken Cook. “Our shopper’s guide to pesticides in produce gives consumers easy

  2. #2 peak

    As RECOGNIZED GENERALLY BY BIBLICAL SCHOLARS, TREATED AS A SCIENCE, DERIVED INDUCTIVELY FROM AN EXEGESIS OF MANY PASSAGES OF SCRIPTURE. By CL

  3. #3 peak

    eighth edition of its Shopper’s Guide to Pesticides in Produce with updated information on 45 popular fruits and vegetables and their total pesticide loads. EWG highlights the worst offenders with its new Dirty Dozen Plus list and the cleanest

  4. #4 peak

    latest USDA tests show we have much more work to do.”⏎EWG researchers analyzed annual pesticide residue tests conducted by the USDA and federal Food and Drug Administration

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. L2:111 suppresses
  2. L2:75 suppresses
  3. L2:49 suppresses
  4. L2:2328 weakly suppresses
  5. L2:1321 weakly suppresses
  6. L2:549 weakly suppresses
  7. L2:2161 weakly excites
  8. L1:2143 weakly suppresses
  9. L2:122 weakly excites
  10. L2:1494 weakly suppresses
  11. L2:2893 weakly excites
  12. L0:1301 weakly suppresses
  13. L0:1224 weakly suppresses
  14. L1:575 weakly excites
  15. L0:1509 weakly excites
  16. L2:1717 weakly excites
  17. L0:178 weakly suppresses
  18. L0:1577 weakly suppresses
  19. L1:491 weakly excites
  20. L2:489 weakly suppresses
  21. L2:493 weakly suppresses
  22. L1:2532 weakly suppresses
  23. L1:125 weakly excites
  24. L1:3004 weakly suppresses

Feeds later layers

  1. L4:873 excites
  2. L19:1495 weakly excites
  3. L8:1795 weakly excites
  4. L16:1837 weakly excites
  5. L5:2285 weakly suppresses
  6. L23:1648 weakly excites
  7. L6:621 weakly excites
  8. L4:2684 weakly suppresses
  9. L21:2371 weakly excites
  10. L4:2771 weakly excites
  11. L5:1801 weakly excites
  12. L8:2874 weakly suppresses
  13. L20:1497 weakly excites
  14. L4:1539 weakly excites
  15. L6:2895 weakly excites
  16. L10:2279 weakly excites
  17. L16:2652 weakly suppresses
  18. L11:1560 weakly suppresses
  19. L5:805 weakly excites
  20. L4:2018 weakly suppresses
  21. L8:1680 weakly excites
  22. L11:3055 weakly excites
  23. L8:1438 weakly suppresses
  24. L4:2566 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 -1.63 and -1.573 values between -0.44 and -0.387 values between -0.38 and -0.3218 values between -0.32 and -0.2655 values between -0.26 and -0.2099 values between -0.20 and -0.14333 values between -0.14 and -0.08765 values between -0.08 and -0.021601 values between -0.02 and 0.04627 values between 0.04 and 0.10239 values between 0.10 and 0.1681 values between 0.16 and 0.2243 values between 0.22 and 0.2823 values between 0.28 and 0.346 values between 0.34 and 0.391 values between 0.39 and 0.452 values between 0.45 and 0.512 values between 0.51 and 0.571 values between 0.57 and 0.631 values between 0.69 and 0.75offnegativepositiveoften

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. 例外
  2. imation
  3. ika
  4. imon
  5. htub
  6. imals
  7. imate
  8. im
  9. (exception
  10. let

Suppresses

  1. 起来了
  2. itest
  3. arta
  4. 针
  5. stakes
  6. usalem
  7. -anchor
  8. altet
  9. ROSS
  10. UGH