Neuron Explorer

Qwen3-0.6B

'still' or 'yet' after a contrast

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

This neuron lights up on words like 'still' or 'yet' when the text has just set up a surprise, such as 'even though X, they still Y'. Turning it up made the model's writing drift to unrelated topics rather than anything about contrast.

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

    hesives,” holding certain proteins to other RNA molecules or to DNA.⏎Now even though the last two jobs I mentioned are not directly related to protein production, they still involve proteins. So is it safe to say that while RNA performs several functions in

  2. #2 peak

    implementing mindful eating into your daily life.⏎Have you ever forgotten to eat lunch because you’re so busy throughout the day? Have you ever eaten a meal and then still felt hungry after? Have you ever mindlessly grazed on a snack? Or

  3. #3 peak

    9 or 10 — they look younger due to malnutrition — and are in grades four or five in their regular government schools. Despite their age, they still don’t know basic math. The toddlers are their siblings, who the girls —

  4. #4 peak

    Ethnic Minorities<|endoftext|>Lily of the valley, or convallaria, is not a true lily like the Easter or Asiatic lily, but is still toxic to dogs, humans and other animals. If your dog chews on the

  5. #5 peak

    -thirds of the way from the center to the edge. At 100,000 light years across, the Milky Way is vast beyond comprehension and yet it’s only one of an estimated 100 billion galaxies in the observable

  6. #6 peak

    for the sake of “looking cooler?”⏎The risk of head injury might be low. But the risk of a car crash is also low, and yet you still wear a seatbelt. The risk of a plane crash is incredibly low, and

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 was just 23, and I was just 23, I was just 23, I was just 23, I was just 23, I was just 23, I was just 2

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

    adult before you fill out any forms, or give out money or credit card details.⏎- Don't accept any offers that seem too good to be true - they probably are.⏎Remember to always be

  2. #2 peak

    have yet to find the perfect thickness, the perfect fill, that they’re certain will reliably support a person’s weight.⏎But with each iteration, they’re getting closer to their goal: a relatively simple, low-cost process for scanning, modeling,

  3. #3 peak

    scale chemicals makers could be fighting for their very survival if carbon reduction targets produce a particularly hostile operating environment. Cost (of production) is not everything, but it is close to it.⏎That fact is brought home particularly in petrochemicals where

  4. #4 peak

    murex shellfish, known as Tyrian purple. This color was, in many cultures of ancient times, reserved for the use of royalty, or at least nobility.⏎It was often attacked by Egypt, besieged by Shalman

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. L12:3057 excites
  2. L12:2794 suppresses
  3. L13:3056 weakly excites
  4. L12:1527 weakly excites
  5. L9:1755 weakly suppresses
  6. L13:26 weakly excites
  7. L13:2169 weakly excites
  8. L12:2773 weakly excites
  9. L12:2717 weakly excites
  10. L11:2712 weakly suppresses
  11. L13:2930 weakly excites
  12. L12:1265 weakly suppresses
  13. L12:1501 weakly suppresses
  14. L13:1039 weakly suppresses
  15. L10:2885 weakly excites
  16. L10:1207 weakly suppresses
  17. L6:590 weakly excites
  18. L13:1859 weakly suppresses
  19. L4:2622 weakly excites
  20. L11:2178 weakly excites
  21. L8:2181 weakly excites
  22. L8:1141 weakly excites
  23. L8:2081 weakly excites
  24. L13:1256 weakly suppresses

Feeds later layers

  1. L15:1462 weakly excites
  2. L23:2579 weakly suppresses
  3. L16:915 weakly excites
  4. L16:696 weakly suppresses
  5. L17:2773 weakly suppresses
  6. L19:718 weakly excites
  7. L27:1397 weakly suppresses "the" in "to the next"
  8. L17:1552 weakly suppresses
  9. L22:2536 weakly suppresses
  10. L27:1593 weakly suppresses
  11. L15:1331 weakly excites
  12. L19:2225 weakly suppresses
  13. L19:212 weakly excites
  14. L18:1866 weakly suppresses
  15. L18:44 weakly excites
  16. L27:2490 weakly suppresses
  17. L25:2203 weakly suppresses
  18. L16:599 weakly suppresses
  19. L23:2110 weakly suppresses
  20. L26:20 weakly excites
  21. L22:774 weakly suppresses
  22. L15:3068 weakly excites
  23. L18:38 weakly excites
  24. L17:1739 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 -1.75 and -1.621 values between -1.49 and -1.354 values between -0.96 and -0.832 values between -0.83 and -0.706 values between -0.70 and -0.567 values between -0.56 and -0.4329 values between -0.43 and -0.30155 values between -0.30 and -0.17800 values between -0.17 and -0.041868 values between -0.04 and 0.10733 values between 0.10 and 0.23216 values between 0.23 and 0.3638 values between 0.36 and 0.4911 values between 0.49 and 0.626 values between 0.62 and 0.762 values between 0.76 and 0.899 values between 0.89 and 1.027 values between 1.02 and 1.152 values between 1.15 and 1.281 values between 1.41 and 1.551 values between 1.55 and 1.681 values between 1.68 and 1.811 values between 1.81 and 1.941 values between 2.34 and 2.472 values between 2.60 and 2.731 values between 2.73 and 2.861 values between 3.00 and 3.131 values between 3.13 and 3.261 values between 3.39 and 3.52offnegativepositiveoften

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. USE
  2. av
  3. because
  4. 불구하고
  5. through
  6. 安全管理
  7. Mission
  8. 实
  9. 仍
  10. FUNCTION

Suppresses

  1. ouched
  2. xhttp
  3. 毗邻
  4. .Reverse
  5. 嗣
  6. ขยาย
  7. urge
  8. imento
  9. utation
  10. ingers