Neuron Explorer

Qwen3-0.6B

'time' in 'for the first time'

A neuron in layer 5 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 fires on the word 'time' in the phrase 'for the first time', and it is quite reliable on that phrase. Turning it up pushed the output toward 'first time' and first-ever events, so the steering matches the examples.

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

    970's when President Nixon visited mainland China for the first time. Today, that may not seem like such a big deal. But back t

  2. #2 peak

    96, the first modern Olympic Games are held in Athens, Greece, with athletes from 14 countries participating.⏎The International Olympic Committee met for the first time in Paris in June 1984 and chose Greece as the site of

  3. #3 peak

    It turns out that robot design could become much more than just a research tool.⏎Team leader Robert J Full of UC Berkeley states, "We showed for the first time that lizards swing their tail up or down to counteract the rotation of their

  4. #4 peak

    accelerators.⏎In 1956, a team of physicists led by Americans Clyde Cowan and Frederick Reines observed the ghostly particles for the first time. For their discovery, Reines shared the 1995 Nobel Prize

  5. #5 peak

    repeated it a few times since then.⏎But there were two new dimensions of the recent flight. One, it carried a Chinese woman astronaut into space for the first time. Secondly and far more significantly, it demonstrated China's capability to conduct docking of

  6. #6 peak

    one of largest cities of the region, also belongs to this number. This name was mentioned for the first time on the earliest copper coins of Buxārā (4th –

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 number of the month is a 1000-year-old tree in the middle of the 2014 cube, the number of the year is a 300-year-old tree, the number of the year is

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

    criteria.<|endoftext|>The haunting image of the fallen soldier, aged just 18, is one of a hundred portraits of those who served in the First World War published on Friday by the Imperial War Museum (IWM).⏎The museum is

  2. #2 peak

    previously done laboriously by hand in hundreds of weavers’ cottages were brought together in a single cotton mill, and the factory was born. The second industrial revolution came in the early 20th century, when Henry Ford mastered the moving

  3. #3 peak

    students learn to tell and write time from analog and digital clocks to the nearest five minutes, using a.m. and p.m.⏎Common Core: 2.MD.7⏎

  4. #4 peak

    961, the 23rd Amendment let District of Columbia residents vote in Federal Elections.⏎In 1964, the 24th Amendment let vote those who could not pay a poll tax.⏎In 196

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. L4:867 suppresses
  2. L4:1034 weakly excites
  3. L1:1993 weakly suppresses
  4. L0:2949 weakly suppresses
  5. L4:1662 weakly suppresses
  6. L0:321 weakly excites
  7. L4:2999 weakly excites
  8. L0:846 weakly suppresses
  9. L4:517 weakly suppresses
  10. L0:2927 weakly excites
  11. L0:998 weakly excites
  12. L2:2437 weakly suppresses
  13. L4:83 weakly excites
  14. L4:1267 weakly excites
  15. L2:650 weakly suppresses
  16. L3:2724 weakly suppresses
  17. L2:2303 weakly suppresses
  18. L4:2201 weakly suppresses
  19. L3:2698 weakly excites
  20. L0:2937 weakly excites
  21. L0:2213 weakly suppresses
  22. L2:82 weakly excites
  23. L1:2482 weakly suppresses
  24. L1:2639 weakly suppresses

Feeds later layers

  1. L6:2213 excites
  2. L18:1378 excites
  3. L11:1663 excites
  4. L6:951 weakly excites
  5. L21:1970 weakly excites
  6. L11:2781 weakly excites
  7. L14:1778 weakly excites
  8. L10:1374 weakly excites
  9. L19:2552 weakly excites
  10. L8:2064 weakly excites
  11. L11:1266 weakly suppresses
  12. L23:2112 weakly suppresses
  13. L18:2837 weakly suppresses
  14. L13:2892 weakly excites
  15. L8:2322 weakly excites
  16. L14:3026 weakly suppresses
  17. L8:1708 weakly excites
  18. L17:3018 weakly suppresses
  19. L16:885 weakly excites
  20. L6:2819 weakly suppresses
  21. L12:1537 weakly excites
  22. L27:793 weakly excites
  23. L18:1125 weakly excites
  24. L13:63 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 -2.54 and -2.401 values between -1.57 and -1.431 values between -1.43 and -1.291 values between -1.29 and -1.151 values between -1.15 and -1.015 values between -0.87 and -0.735 values between -0.73 and -0.606 values between -0.60 and -0.4643 values between -0.46 and -0.32193 values between -0.32 and -0.18990 values between -0.18 and -0.042021 values between -0.04 and 0.10501 values between 0.10 and 0.2496 values between 0.24 and 0.3815 values between 0.38 and 0.527 values between 0.52 and 0.657 values between 0.65 and 0.792 values between 0.79 and 0.932 values between 0.93 and 1.072 values between 1.07 and 1.211 values between 1.21 and 1.352 values between 1.63 and 1.761 values between 1.76 and 1.901 values between 1.90 and 2.042 values between 2.46 and 2.601 values between 2.88 and 3.01offnegativepositiveoften

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. ERICA
  2. 机会
  3. 等到
  4. les
  5. blush
  6. acles
  7. 美的
  8. orque
  9. sourceMappingURL
  10. egl

Suppresses

  1. jąc
  2. alyze
  3. kip
  4. anoia
  5. annya
  6. úde
  7. télé
  8. 隍
  9. 릭
  10. umen