Neuron Explorer

Qwen3-0.6B

"the" in mid-sentence, weak

A neuron in layer 20 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 weakly responds to the word "the" in the middle of a sentence, especially after a verb or preposition. Its strongest opposite reactions come on words like "over", "and" and "than" in comparisons such as "one over the other". The steered text was loosely about lists and comparisons, but not clearly.

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

    you need to ask before you can identify an individual?⏎When you’re looking at twenty-four candidates, the worst case scenario is twenty-two (assuming it’s not the person on your card) if you just ask “Is your person [name]

  2. #2 peak

    otherwise, the two will ot stick together. If it is the job of a protein to catalyze a chemical reaction, then the shape of

  3. #3 peak

    second bottle to be the fake.⏎In practice, there are some limitations to a manual application of this geometric

  4. #4 peak

    : Associated Press⏎About one in seven pregnant women get flu shots even though health experts encourage them to be the first in line every flu season.⏎Many obstetricians don't vaccinate and

  5. #5 peak

    stocks that have dividend yields of 15% or more. To a beginner investor, stocks like this may seem like a bargain. However, it is often the case that the dividends for these high yielding companies are unsustainable and only high because the

  6. #6 peak

    stomach cancer than white adults.⏎- In the poorest U.S. counties, men with colorectal cancer have a 35% higher death rate than men in the most affluent counties.⏎- Bisexual women are 70% more likely to

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 fact that the word the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the the

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

    ivalent’ non-PWM fan, but non-PWM fans can still run slow enough. For the average consumer I can’t think of much reason to prefer one over the other.⏎Can you control RGB colors

  2. #2 peak

    pace. To switch between modelspace and paperspace, view the tabs located at the bottom of the screen. One tab is labeled ‘modelspace’ and the other tabs are labeled either as ‘sheet’ or ‘layout’. The ‘

  3. #3 peak

    both good options when it comes to charging your battery efficiently. When differentiating PWM and MPPT, the discussion is not going to be about which one is better than the other.⏎What should be determined when choosing between PWM and MP

  4. #4 peak

    .⏎Counseling Point for Anti-Platelets:⏎Aspirin and warfarin are sometimes prescribed together, and this is because one is anti platelet and the other is anticoagulant. Please see the previous anticoagulant

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. L19:2606 strongly excites
  2. L19:1933 strongly excites
  3. L19:2013 excites
  4. L18:2251 suppresses
  5. L18:2780 suppresses
  6. L17:2968 excites
  7. L18:2605 suppresses
  8. L19:2745 excites
  9. L19:2789 excites
  10. L18:2915 suppresses
  11. L19:3028 suppresses
  12. L17:2957 excites
  13. L17:1724 suppresses
  14. L18:2953 excites
  15. L19:2283 suppresses
  16. L19:2418 excites
  17. L19:2587 excites
  18. L18:1458 suppresses
  19. L16:1510 weakly excites
  20. L4:854 weakly suppresses
  21. L18:2241 weakly excites
  22. L19:2703 weakly excites
  23. L18:2256 weakly excites
  24. L15:130 weakly suppresses

Feeds later layers

  1. L26:507 strongly suppresses
  2. L27:127 strongly excites
  3. L21:863 strongly suppresses
  4. L27:399 strongly suppresses
  5. L23:94 excites
  6. L24:49 excites
  7. L21:758 suppresses
  8. L26:535 excites
  9. L22:199 suppresses
  10. L22:171 suppresses
  11. L27:1464 excites
  12. L23:371 excites
  13. L25:2531 suppresses
  14. L27:1382 weakly suppresses
  15. L26:1200 weakly suppresses
  16. L27:2537 weakly excites
  17. L27:994 weakly excites
  18. L25:2497 weakly suppresses
  19. L21:1232 weakly suppresses
  20. L27:121 weakly excites
  21. L25:1796 weakly suppresses
  22. L21:488 weakly suppresses
  23. L27:25 weakly suppresses
  24. L26:2278 weakly 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.

1 values between -18.78 and -18.111 values between -5.39 and -4.721 values between -4.72 and -4.052 values between -4.05 and -3.381 values between -3.38 and -2.723 values between -2.72 and -2.0514 values between -2.05 and -1.3867 values between -1.38 and -0.71765 values between -0.71 and -0.042579 values between -0.04 and 0.63395 values between 0.63 and 1.3027 values between 1.30 and 1.9716 values between 1.97 and 2.6414 values between 2.64 and 3.317 values between 3.31 and 3.982 values between 3.98 and 4.655 values between 4.65 and 5.324 values between 5.32 and 5.991 values between 5.99 and 6.663 values between 7.33 and 8.00offnegativepositiveoften

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. others
  2. 除此之外
  3. Others
  4. 其他
  5. others
  6. OTHER
  7. another
  8. Others
  9. another
  10. other

Suppresses

  1. the
  2. the
  3. The
  4. _the
  5. The
  6. THE
  7. -the
  8. →the
  9. →The
  10. ethe