Neuron Explorer

Qwen3-0.6B

'said' in 'is said to'

A neuron in layer 1 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 'said' in phrases like 'is said to be', where a claim is passed on as a reputation. Turning it up did not produce clear output that matches this pattern.

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

    anan. There is a temple in the Tulunadu known by the name Yathugiri which is situated on a mountainous region. This temple is said to have 9 sacred water. This place is also known by the name Ya

  2. #2 peak

    politically or socially motivated purpose. The individual who performs an act of hacktivism is said to be a hacktivist.⏎Form Of Hacktivism :⏎Wikipedia is

  3. #3 peak

    risk factors also crop up as indicators in other diseases as well<|endoftext|>You are attending a highly exclusive seminar run by the famous artist known only as Leo.⏎Leo is said to be highly influential in the art world. In fact,

  4. #4 peak

    Mahadev Temple is said to be of ancient times. According to local legend, the temple was built by

  5. #5 peak

    which can then be glued to a different kind of wood to create a solid walnut appearance without the cost of actual solid walnut wood. Walnut heartwood in particular is said to be unique, as it appears to be deeply

  6. #6 peak

    that students have two important models of growth in the real world: linear and exponential. Have students compare and contrast them.⏎- The terms of an arithmetic sequence are said to grow “linearly.” The terms of a geometric sequence are said to grow

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 thinking about the relationship between the two things, the two things that are not only the two things that are not only the two things that are not only the two things. I have been thinking about the relationship between the two things

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

    its employee, the bus driver.⏎When governmental immunity applies,<|endoftext|>Actinopteri (ray-finned fishes) > Cichliformes⏎(Cichlids, convict blennies) > Cichlidae⏎(Cich

  2. #2 peak

    brata (vertebrates) > Gnathostomata (jawed⏎vertebrates) > Teleostomi (teleost fish) > Osteichthyes (bony fish) > Class:⏎fish) > Stegocephalia

  3. #3 peak

    2: Measurement Adapter⏎Fig.3: Measurement Adapter Schematics⏎The Capacitors used are Mica (Glimmer) obtained from www.reichelt.de⏎With a jumper either the short or one of the series capacitors

  4. #4 peak

    large fish some 1.5 meters long. Its entire body is covered with scales, reminiscent of armor plating. It belongs to the Osteoichthyes class of bony fishes, of

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. L0:1301 weakly suppresses
  2. L0:1224 weakly suppresses
  3. L0:178 weakly suppresses
  4. L0:1021 weakly excites
  5. L0:1360 weakly excites
  6. L0:451 weakly excites
  7. L0:732 weakly excites
  8. L0:120 weakly excites
  9. L0:84 weakly suppresses
  10. L0:143 weakly suppresses
  11. L0:851 weakly suppresses
  12. L0:30 weakly suppresses
  13. L0:2304 weakly excites
  14. L0:3070 weakly excites
  15. L0:1245 weakly excites
  16. L0:866 weakly suppresses
  17. L0:2278 weakly excites
  18. L0:116 weakly suppresses
  19. L0:1859 weakly suppresses
  20. L0:139 weakly excites
  21. L0:197 weakly suppresses
  22. L0:418 weakly suppresses
  23. L0:1831 weakly excites
  24. L0:1159 weakly excites

Feeds later layers

  1. L4:1118 weakly excites
  2. L21:2820 weakly excites
  3. L14:409 weakly excites
  4. L8:1432 weakly excites
  5. L6:738 weakly excites
  6. L2:247 weakly excites
  7. L7:2904 weakly excites
  8. L11:1781 weakly suppresses
  9. L19:925 weakly excites
  10. L4:1864 weakly excites
  11. L27:146 weakly suppresses
  12. L16:1008 weakly suppresses
  13. L7:213 weakly excites
  14. L5:2362 weakly suppresses
  15. L17:1551 weakly excites
  16. L2:2367 weakly suppresses
  17. L11:1645 weakly excites
  18. L18:2991 weakly suppresses
  19. L8:1812 weakly suppresses
  20. L9:1412 weakly excites the word "be" after modals and "to"
  21. L22:933 weakly suppresses
  22. L10:2744 weakly excites
  23. L10:605 weakly suppresses
  24. L5:1594 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 -1.17 and -1.044 values between -0.53 and -0.4032 values between -0.40 and -0.28320 values between -0.28 and -0.151656 values between -0.15 and -0.021596 values between -0.02 and 0.10207 values between 0.10 and 0.2341 values between 0.23 and 0.3612 values between 0.36 and 0.498 values between 0.49 and 0.618 values between 0.61 and 0.741 values between 0.74 and 0.877 values between 0.87 and 1.003 values between 1.00 and 1.122 values between 1.12 and 1.252 values between 1.25 and 1.382 values between 1.38 and 1.501 values between 2.01 and 2.142 values between 2.27 and 2.401 values between 2.40 and 2.521 values between 2.65 and 2.781 values between 3.80 and 3.92offnegativepositiveoften

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. oton
  2. ульт
  3. antly
  4. 烏
  5. нд
  6. nt
  7. äch
  8. ión
  9. 对此
  10. oreal

Suppresses

  1. amenti
  2. handleClose
  3. icina
  4. 廋
  5. лежа
  6. downtown
  7. atted
  8. 嫌疑
  9. Cooperation
  10. себе