Neuron Explorer

Qwen3-0.6B

'gl' starting words like glutes

A neuron in layer 26 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

It fires on the 'gl' sound at the start of a word split into pieces, such as 'glutes', 'glands', 'glaucoma', 'glia' and 'glint'. It turns down on words about prices and costs, like 'price' and 'cost'. Turning it up made the model repeat the word 'sound' or drift into unrelated lines.

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

    also strengthen your upper body.⏎Move More Muscles⏎Most of the muscles you use in running are in your lower body: your quads, hamstrings, glutes, hip flexors and calves

  2. #2 peak

    and lower-body muscles serve as your primary movers, and you strengthen many more muscles than when running. In addition to your calves, hamstrings, quadriceps and glutes, you strengthen your abs and erector spinae in your core,

  3. #3 peak

    -white discharge. There may be spread on to the perineum and into the groins. In Candida balanitis, tiny papules develop on the glans penis, evolve as white pustules or vesicles and rupture, leaving

  4. #4 peak

    wash your hands often, and avoid touching your eyes.⏎This is especially true if you have had any eye surgery such as LASIK, cataract surgery or glaucoma surgery, since your eyes may be more prone

  5. #5 peak

    85 percent.⏎The group name for white matter cells, glia—derived from the world ‘glue’—reflects their lowly

  6. #6 peak

    claimed to have been knocked out by a glancing blow, and said that when he came to, he found that the cook

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: sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound sound

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

    is the actual rate of price inflation. They don’t use the CPI ratio put out by the Census Bureau. They use what is called a “chained” price deflator from the GDP report which means they take prices as they were in

  2. #2 peak

    deviation of the seven-month change in the inflation series for that item since 2000. (We go back only to 2000 because price indexes for some of the twenty-three items are not available before then.) Bars of

  3. #3 peak

    a reduction in volume per plant. Thus, total manufacturing costs have been increased about a cent a pound, compared with the same months a year ago.⏎Using typical cost figures supplied us from various types of creameries, it is evidence that it costs

  4. #4 peak

    market movement. It is one of the major components of the forward price index methodology. Variable cost index also known as VAT, is often confused with the expression “cost of good sale” which should not be confused as the same thing. The main

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. L21:873 suppresses
  2. L3:1824 suppresses
  3. L9:1061 weakly excites
  4. L9:991 weakly suppresses
  5. L4:2237 weakly excites
  6. L10:1119 weakly excites
  7. L6:772 weakly excites
  8. L10:1042 weakly suppresses
  9. L10:1286 weakly suppresses
  10. L6:401 weakly excites
  11. L22:113 weakly excites
  12. L21:2455 weakly suppresses
  13. L16:126 weakly excites
  14. L15:1595 weakly excites
  15. L8:1596 weakly suppresses
  16. L21:1588 weakly excites
  17. L10:2195 weakly suppresses
  18. L24:151 weakly excites
  19. L7:2823 weakly excites
  20. L3:1346 weakly suppresses
  21. L19:171 weakly suppresses
  22. L16:2287 weakly suppresses
  23. L6:1978 weakly suppresses
  24. L6:2350 weakly suppresses

Feeds later layers

  1. L27:1613 weakly suppresses
  2. L27:772 weakly suppresses
  3. L27:2230 weakly excites
  4. L27:1331 weakly excites
  5. L27:1442 weakly suppresses
  6. L27:1657 weakly excites
  7. L27:1997 weakly suppresses
  8. L27:1904 weakly suppresses
  9. L27:2466 weakly suppresses
  10. L27:2510 weakly suppresses
  11. L27:2004 weakly excites the 's' after a hyphen
  12. L27:491 weakly suppresses
  13. L27:2214 weakly suppresses
  14. L27:1336 weakly suppresses
  15. L27:2999 weakly suppresses
  16. L27:1162 weakly excites
  17. L27:2878 weakly suppresses
  18. L27:1730 weakly suppresses
  19. L27:1688 weakly suppresses
  20. L27:991 weakly excites
  21. L27:369 weakly excites
  22. L27:2836 weakly excites
  23. L27:2416 weakly suppresses
  24. L27:2558 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 -17.67 and -16.601 values between -15.54 and -14.471 values between -13.40 and -12.331 values between -12.33 and -11.264 values between -11.26 and -10.194 values between -10.19 and -9.134 values between -9.13 and -8.068 values between -8.06 and -6.999 values between -6.99 and -5.9214 values between -5.92 and -4.8525 values between -4.85 and -3.7827 values between -3.78 and -2.7271 values between -2.72 and -1.65400 values between -1.65 and -0.582966 values between -0.58 and 0.49283 values between 0.49 and 1.5640 values between 1.56 and 2.6322 values between 2.63 and 3.7011 values between 3.70 and 4.766 values between 4.76 and 5.833 values between 5.83 and 6.903 values between 6.90 and 7.971 values between 12.24 and 13.311 values between 14.38 and 15.451 values between 19.72 and 20.791 values between 23.99 and 25.06offnegativepositiveoften

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. imulation
  2. acial
  3. odata
  4. кра
  5. _SRV
  6. sábado
  7. ibrate
  8. クラ
  9. RVA
  10. bearer

Suppresses

  1. izados
  2. plat
  3. sound
  4. ertainment
  5. Comput
  6. ological
  7. chain
  8. ennis
  9. maxlen
  10. ��