Neuron Explorer

Qwen3-0.6B

'as' in 'as far as' and 'as many as'

A neuron in layer 24 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 reacts to the word "as" in comparison phrases such as "as far as", "as early as" or "as many as". Its negative side fires on words like "over" in "overkill" and "too" in "too light-handed", which are less clearly related. When turned up or down, the model's writing became repetitive or off topic, so steering gives no clear sign.

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

    anches, previously the Bishops of Coutances. Standing 80 metres (295 ft) tall, it dominates the town and can be seen from as far away as the island of Jersey. It is a classic example of the Gothic

  2. #2 peak

    they don't do as well as younger people on complex memory or learning tests. Scientists have found, though, that given enough time, healthy older people can do as well as younger people do on these tests. In fact, as they age,

  3. #3 peak

    besides an exposition of Dioscorides, a Natural History dealing with all the plants known to Mattioli.<|endoftext|>The History of the hall can be traced to as far back as the Domesday Survey. With little knowledge of the hall prior

  4. #4 peak

    possible to preempt<|endoftext|>Food for thought⏎Sit down for a meal at the table, and you'll probably find foods from a half-dozen countries spread across nearly as many continents, on your plate.⏎Supermarket shelves today are likely to be stocked

  5. #5 peak

    of HCV infection is dependant upon on geography, alcohol use, viral characteristics ( different genetic types, viral load ), co-infection with other viruses and some as yet unidentified factors.⏎After exposure to the virus, detectable viral genetic material called

  6. #6 peak

    -dial or a garden bed or the residence itself, or a brook falling down the sward between trees and bushes and clumpy growths. There are as many forms and kinds of gardens as there are persons who have gardens […].”

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: how to use a perking on a perking on a perking on a perking... perking perking... perking on a perking perking... on a perking on a perking... on a perking perking

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

    the Clean Water Act, it just brings a new truckload of restrictions for the people who do most to protect our water.⏎The Clean Water Restoration Act is regulatory overkill. It is written<|endoftext|>Feeling anxious about stressful events is a normal response

  2. #2 peak

    . To guarantee Southerners their much-desired states' rights, the federal government had no authority to levy protective tariffs, make internal improvements, or overrule state court decisions, while states had the right to sustain their own armies and

  3. #3 peak

    are being applied.⏎“If fracking is not done well it can have significant environmental impacts including polluting water and triggering earthquakes.⏎“I am also concerned that regulation may be too light-handed, particularly if fracking opens the door to a large-scale and widespread oil

  4. #4 peak

    y often criticized as overabundant and extravagant. Mercedes Maroto Camino (1995

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. L13:2250 excites
  2. L21:2777 excites
  3. L22:151 weakly suppresses
  4. L16:628 weakly suppresses
  5. L18:2374 weakly suppresses
  6. L9:2431 weakly excites
  7. L10:103 weakly excites
  8. L19:569 weakly excites
  9. L2:929 weakly suppresses
  10. L23:206 weakly suppresses
  11. L7:795 weakly suppresses
  12. L8:1924 weakly suppresses
  13. L1:2164 weakly suppresses
  14. L11:2999 weakly suppresses
  15. L7:1754 weakly excites
  16. L4:1931 weakly excites
  17. L10:644 weakly suppresses
  18. L23:3034 weakly excites
  19. L18:102 weakly suppresses
  20. L9:2760 weakly suppresses
  21. L8:893 weakly excites
  22. L7:1938 weakly suppresses
  23. L4:1698 weakly excites
  24. L22:1723 weakly suppresses

Feeds later layers

  1. L26:1471 weakly excites
  2. L27:2822 weakly excites
  3. L27:1233 weakly excites
  4. L25:2727 weakly excites
  5. L26:268 weakly suppresses
  6. L26:890 weakly suppresses
  7. L26:761 weakly suppresses
  8. L27:991 weakly suppresses
  9. L25:892 weakly excites
  10. L26:1810 weakly excites
  11. L26:655 weakly suppresses
  12. L27:2653 weakly excites
  13. L25:1154 weakly excites
  14. L27:837 weakly excites
  15. L27:1458 weakly excites
  16. L27:1437 weakly excites
  17. L27:1821 weakly excites
  18. L25:1357 weakly suppresses
  19. L25:2618 weakly suppresses
  20. L25:2356 weakly excites
  21. L25:714 weakly suppresses
  22. L25:1863 weakly excites
  23. L27:888 weakly excites
  24. L26:2874 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 -33.97 and -32.452 values between -20.28 and -18.762 values between -17.24 and -15.721 values between -14.20 and -12.672 values between -12.67 and -11.152 values between -11.15 and -9.633 values between -9.63 and -8.111 values between -8.11 and -6.594 values between -6.59 and -5.0712 values between -5.07 and -3.5519 values between -3.55 and -2.03462 values between -2.03 and -0.513276 values between -0.51 and 1.0281 values between 1.02 and 2.5421 values between 2.54 and 4.0610 values between 4.06 and 5.583 values between 5.58 and 7.102 values between 7.10 and 8.621 values between 8.62 and 10.141 values between 13.19 and 14.711 values between 17.75 and 19.271 values between 25.35 and 26.88offnegativepositiveoften

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. PLY
  2. etus
  3. urses
  4. Cena
  5. ряд
  6. oted
  7. tern
  8. 酡
  9. ENCY
  10. voke

Suppresses

  1. bs
  2. on
  3. per
  4. quee
  5. len
  6. amb
  7. alsa
  8. Ste
  9. king
  10. Ste