Neuron Explorer

Qwen3-0.6B

'back' after a verb or noun

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 fires on the word 'back' when it follows a verb or noun, as in 'backup', 'built back' or 'grown back'. Turning it up did not clearly make the writing mention 'back', so the steering evidence is mixed.

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

    employers may keep the SDSs in a binder or on computers as long as the employees have immediate access to the information without leaving their work area when needed and a back-up is available for rapid access to the SDS in the

  2. #2 peak

    959, but the name peaked in popularity in 1973, in Nebraska.⏎Martin has an old-fashioned ring to it, and the numbers back up that impression--the name peaked in 1964 and was especially

  3. #3 peak

    on a site's soil conditions. Root systems usually are shallower in heavy clay soils.⏎In some situations, it can be detrimental to amend or replace heavy clay backfill soil, because it can trap water around the tree's roots and also can

  4. #4 peak

    received information on their change in heart rates and blood pressure, I added this to the spreadsheet.⏎As this was happening the data was also being displayed in a graph back in our classroom about 100m away. On completion of the laboratory

  5. #5 peak

    each site, fish, macro-invertebrates, and benthic composition information is then quantified following standardized protocols. By relating the data collected in the field back to the habitat maps and bathymetric models, BB is able to model and

  6. #6 peak

    later, man would eventually take those first steps on the moon, meaning the United States had beaten Russia in the space race, but it also led to new changes back down on Earth.⏎“A lot of the things we developed to engage in the space

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 small family (family worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth worth

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

    Dry mouth⏎- Increased thirst⏎- Eyes that appear sunken⏎- Skin changes -- for example, if you touch or squeeze the skin, it doesn't bounce back the way it usually does⏎- Urinating less often or having dark yellow

  2. #2 peak

    igious affiliation within Cameroon boasts a whopping 70% Christian. 21% is of the Islamic faith, and 6% still maintain indigenous belief systems dating back to the cultural inception thousands of years ago. Islamic believers are concentrated in the

  3. #3 peak

    the actual practice (conducted in the arena).⏎These two continually learn from each other and evolve.⏎The beginnings of the science⏎Although the theories and logic used date back to Aristotle, the modern application began in the 1960′

  4. #4 peak

    to heat pumps and solar thermal, we must look at the period needed to cover their upfront cost through the savings on energy bills, which is known as the “payback time”.⏎With the existing incentives and prices, an average middle-income family

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. L21:2530 strongly suppresses
  2. L21:2076 suppresses
  3. L21:1525 suppresses
  4. L14:2160 excites
  5. L23:2231 excites
  6. L17:1898 suppresses
  7. L21:2998 suppresses
  8. L19:3036 excites
  9. L18:3044 suppresses
  10. L7:196 excites
  11. L15:1563 excites
  12. L23:2861 excites
  13. L22:2937 weakly excites
  14. L22:2615 weakly suppresses
  15. L23:1195 weakly suppresses
  16. L19:1951 weakly suppresses
  17. L17:2799 weakly suppresses
  18. L19:1871 weakly suppresses
  19. L11:1688 weakly suppresses
  20. L5:2709 weakly excites
  21. L23:364 weakly excites
  22. L23:1182 weakly suppresses
  23. L7:2996 weakly suppresses
  24. L22:1872 weakly excites

Feeds later layers

  1. L27:1173 strongly suppresses
  2. L26:2189 suppresses
  3. L26:1281 suppresses
  4. L25:2242 weakly suppresses
  5. L25:1799 weakly suppresses
  6. L25:2159 weakly suppresses
  7. L26:2629 weakly suppresses
  8. L27:1179 weakly suppresses
  9. L27:2930 weakly suppresses
  10. L26:499 weakly suppresses
  11. L25:2679 weakly suppresses
  12. L27:928 weakly suppresses
  13. L25:2304 weakly suppresses
  14. L25:1660 weakly suppresses
  15. L26:2010 weakly suppresses
  16. L27:1686 weakly suppresses
  17. L26:1637 weakly excites
  18. L27:1820 weakly suppresses
  19. L25:2062 weakly excites
  20. L27:2032 weakly suppresses
  21. L27:798 weakly excites
  22. L26:3018 weakly excites
  23. L25:1012 weakly suppresses
  24. L25:275 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 -20.77 and -19.661 values between -15.23 and -14.122 values between -13.02 and -11.912 values between -11.91 and -10.804 values between -10.80 and -9.709 values between -8.59 and -7.486 values between -7.48 and -6.3710 values between -6.37 and -5.2715 values between -5.27 and -4.1620 values between -4.16 and -3.0538 values between -3.05 and -1.95150 values between -1.95 and -0.842981 values between -0.84 and 0.27487 values between 0.27 and 1.3897 values between 1.38 and 2.4829 values between 2.48 and 3.5917 values between 3.59 and 4.7015 values between 4.70 and 5.806 values between 5.80 and 6.919 values between 6.91 and 8.021 values between 8.02 and 9.121 values between 9.12 and 10.231 values between 10.23 and 11.341 values between 11.34 and 12.451 values between 12.45 and 13.551 values between 14.66 and 15.771 values between 15.77 and 16.872 values between 22.41 and 23.52offnegativepositiveoften

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. 门户
  2. idable
  3. umu
  4. умент
  5. porto
  6. 缢
  7. uisine
  8. umen
  9. senal
  10. ulpt

Suppresses

  1. back
  2. back
  3. Back
  4. Back
  5. -back
  6. (back
  7. →back
  8. _back
  9. BACK
  10. backs