Neuron Explorer

Qwen3-0.6B

'called' after 'so'

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 fires on the word 'called' when it follows 'so', as in 'so-called'. Turning it up did not change the replies in any way that matches this pattern.

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

    can just concentrate on learning how to make machines obey digital commands. The engineering grunt comes after the intellectual workout.⏎You might think it strange that the makers of so-called educational computers are so busy trying to maintain their market share that they failed to see

  2. #2 peak

    through a buy-an-acre program. It is not uncommon for loggers, oil and gas companies, cattle ranchers, and miners to illegally extract resources from so-called “protected” areas.⏎Traditional forest communities are often the best stewards of the

  3. #3 peak

    , are of some specified fixed number of bits. For example, Java collections use 32-bit hash codes.⏎In cryptography applications, we often need a so-called secure hash function. Secure hash

  4. #4 peak

    location and length of horizontal transfers (HTs) are known for very few species. The usage of short oligonucleotides in a sequence (the so-called genomic signature) has been shown to be species-specific even in

  5. #5 peak

    space?" The variety of answers illustrates the function of fundamental analogies or models in physics as well as the role of so-called unobservable entities. Forces and Fields presents an in-depth look at the science of

  6. #6 peak

    light, the researchers have shown that the signaling pathways, which are activated by the receptors play a crucial role in controlling movement.⏎Unlike conventional methods, with the so-called optogenetics, the researchers were able to target one cell type. "We

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 problem of a girl with a number, and the number of the problem. What did you think? It's a problem about a girl. But here is a problem. Let's go through this. So, etc. So, and

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

    uncovered that Patagonia was filling jackets using suppliers who plucked down from birds that were still alive and force-fed so their livers could be used as foie gras. After the Patagonia incident, many other brands underwent investigation and uncovered

  2. #2 peak

    rebuilding the economy. A letter from King Bela IV refers to Jews in the cities of Pressburg (Bratislava), Senica, Trnava, Pezinok, Nitra, and Trencin. In 1

  3. #3 peak

    , 1971, p. 37; Boyce, 1977, p. 42).⏎In other Middle

  4. #4 peak

    Ewan Boyce and later the property of her husband Captain William Boyce<|endoftext|>A Point of View: Making friends the shared currency way⏎- 2

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. L1:92 weakly excites
  2. L2:1206 weakly suppresses
  3. L1:2801 weakly excites
  4. L2:1692 weakly suppresses
  5. L1:2056 weakly suppresses
  6. L1:2054 weakly excites
  7. L0:2456 weakly excites
  8. L1:2679 weakly suppresses
  9. L2:506 weakly suppresses
  10. L1:1979 weakly excites
  11. L1:87 weakly suppresses
  12. L2:1955 weakly suppresses
  13. L0:980 weakly suppresses
  14. L2:1899 weakly excites
  15. L1:1963 weakly suppresses
  16. L2:2801 weakly suppresses
  17. L1:2938 weakly suppresses
  18. L1:571 weakly excites
  19. L2:493 weakly suppresses
  20. L0:1259 weakly suppresses
  21. L2:2397 weakly excites
  22. L0:2062 weakly excites
  23. L1:97 weakly excites full stop ending a heading or label
  24. L2:1765 weakly suppresses

Feeds later layers

  1. L13:1666 weakly excites
  2. L5:2799 weakly excites
  3. L18:977 weakly excites
  4. L9:968 weakly excites
  5. L15:1581 weakly excites
  6. L12:1141 weakly suppresses
  7. L4:577 weakly suppresses
  8. L19:1208 weakly suppresses
  9. L23:1484 weakly excites
  10. L12:811 weakly excites
  11. L9:2691 weakly excites
  12. L4:2471 weakly suppresses
  13. L14:675 weakly suppresses
  14. L10:2446 weakly excites
  15. L17:294 weakly suppresses
  16. L4:1125 weakly suppresses
  17. L7:1745 weakly suppresses
  18. L24:779 weakly suppresses
  19. L5:2289 weakly suppresses
  20. L19:298 weakly excites
  21. L8:394 weakly excites
  22. L4:146 weakly excites
  23. L15:58 weakly excites
  24. L6:1425 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 -1.73 and -1.661 values between -1.53 and -1.472 values between -1.02 and -0.961 values between -0.96 and -0.902 values between -0.51 and -0.456 values between -0.45 and -0.3910 values between -0.39 and -0.3236 values between -0.32 and -0.2681 values between -0.26 and -0.20197 values between -0.20 and -0.13516 values between -0.13 and -0.071022 values between -0.07 and -0.001123 values between -0.00 and 0.06555 values between 0.06 and 0.12199 values between 0.12 and 0.19104 values between 0.19 and 0.2539 values between 0.25 and 0.317 values between 0.31 and 0.384 values between 0.38 and 0.441 values between 0.51 and 0.571 values between 0.76 and 0.82offnegativepositiveoften

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. thin
  2. 闩
  3. spécial
  4. ector
  5. darn
  6. üc
  7. 技术支持
  8. ncpy
  9. 主力
  10. şa

Suppresses

  1. LLU
  2. чив
  3. anker
  4. anken
  5. ante
  6. явление
  7. Kan
  8. ente
  9. kan
  10. ética