Neuron Explorer

Qwen3-0.6B

'than' after comparative adjectives

A neuron in layer 16 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 'than' when it follows a comparison, such as 'happier than' or 'higher than'. On the negative side it fires on 'to' or 'about' after verbs and phrases like 'when it comes to'. Turning it up produced repetitive, looping text.

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

    ers Fighting Drowsiness⏎Conclusion. The Employers' Dreams⏎One thing is absolutely certain in America: the quality and quantity of sleep obtained is substantially less than the quality and quantity that are needed. Over the past century, we have reduced

  2. #2 peak

    to take his perspective.⏎What does this mean for the kiddos of authoritative parents? They grow up knowing how to make their own decisions, and are generally happier than kids who have parents that choose other styles. Does this mean that 10

  3. #3 peak

    ," explains Wrigley.⏎Infants are by far the most vulnerable children in care. Their fatality rate from both accidents and violence is nearly seven times higher than that of children from one to four. Equally striking are differences in infant fat

  4. #4 peak

    others increase their happiness, it also increases our own. Studies amongst different cultures and ages have consistently found that individuals who spend money on others report higher levels of happiness than those who spend money on themselves. Researchers suggest that the reward we experience from helping

  5. #5 peak

    cut the emission of carbon dioxide, methane, and other greenhouse gases.⏎It is clearer now than ever that the U.S. and other developed nations can and must act quickly and

  6. #6 peak

    .S. counties in which new Wal-Mart stores were built between 1987 and 1998 experienced a significantly smaller reduction in their poverty rates than those counties that did not add new Wal-Mart stores.⏎Overall, the portion of

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 a 2000s, 2010s, 2000s, 2000s, 2000s, 2000s, 20

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

    3” below the surface) to deep bands at 6” or lower. Deep applications tend to be most beneficial in dry soil conditions. When it comes to timing, Vyn recommends banding P and K in the fall and spring applications

  2. #2 peak

    , I'm tempted to butt in and ask if you're talking about the Mars rover Curiosity. She's really there! On Mars! Right now

  3. #3 peak

    are obligated for few formalities. They need to ensure that the professional details, whether relating to the employee or the employer, are to<|endoftext|>It takes a while before one can

  4. #4 peak

    is, and how little equipment is needed. I’m not that smart; if I can do it, so can you!⏎To be clear, I’m talking about processing (developing, not printing) black and white (35mm or

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. L10:272 suppresses
  2. L15:517 excites
  3. L14:630 weakly suppresses
  4. L3:2116 weakly excites
  5. L12:1925 weakly excites
  6. L10:2035 weakly suppresses
  7. L4:462 weakly excites
  8. L5:2671 weakly excites
  9. L15:2995 weakly suppresses
  10. L6:1293 weakly excites
  11. L6:1880 weakly suppresses "by" introducing a deadline or means
  12. L9:792 weakly suppresses
  13. L10:155 weakly suppresses
  14. L10:2098 weakly suppresses
  15. L14:2150 weakly suppresses
  16. L6:1787 weakly excites
  17. L6:914 weakly suppresses
  18. L1:469 weakly excites
  19. L3:1292 weakly suppresses
  20. L4:3054 weakly excites
  21. L13:2166 weakly suppresses
  22. L7:138 weakly excites 'to' after verbs like prove or time
  23. L10:132 weakly excites
  24. L8:45 weakly suppresses

Feeds later layers

  1. L17:2192 excites
  2. L17:865 suppresses
  3. L19:2397 suppresses
  4. L18:606 suppresses
  5. L21:225 excites
  6. L18:758 suppresses
  7. L17:1592 excites
  8. L18:134 excites
  9. L20:271 excites
  10. L18:370 suppresses
  11. L18:2093 excites
  12. L24:150 weakly suppresses
  13. L19:1434 weakly suppresses
  14. L20:482 weakly suppresses
  15. L18:188 weakly suppresses
  16. L23:24 weakly excites
  17. L24:297 weakly excites
  18. L17:2877 weakly excites
  19. L18:2271 weakly suppresses
  20. L19:829 weakly suppresses
  21. L19:384 weakly suppresses
  22. L26:731 weakly excites
  23. L18:2410 weakly excites
  24. L19:3054 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 -3.29 and -3.101 values between -3.10 and -2.922 values between -2.92 and -2.741 values between -2.74 and -2.551 values between -2.37 and -2.181 values between -2.18 and -2.003 values between -2.00 and -1.811 values between -1.81 and -1.631 values between -1.63 and -1.444 values between -1.44 and -1.269 values between -1.26 and -1.0714 values between -1.07 and -0.8921 values between -0.89 and -0.7027 values between -0.70 and -0.52114 values between -0.52 and -0.33515 values between -0.33 and -0.151920 values between -0.15 and 0.04846 values between 0.04 and 0.22210 values between 0.22 and 0.4073 values between 0.40 and 0.5938 values between 0.59 and 0.7726 values between 0.77 and 0.9628 values between 0.96 and 1.1419 values between 1.14 and 1.339 values between 1.33 and 1.515 values between 1.51 and 1.703 values between 1.70 and 1.883 values between 1.88 and 2.075 values between 2.07 and 2.252 values between 2.25 and 2.441 values between 2.44 and 2.621 values between 2.62 and 2.811 values between 3.17 and 3.361 values between 3.54 and 3.731 values between 3.91 and 4.10offnegativepositiveoften

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. dú
  2. )[:
  3. рад
  4. ),(
  5. 痕
  6. /sbin
  7. 涌现出
  8. ((⏎
  9. 涉
  10. وى

Suppresses

  1. anship
  2. elder
  3. elevation
  4. Lesser
  5. Python
  6. snapping
  7. .scala
  8. lies
  9. !_
  10. Moose