Neuron Explorer

Qwen3-0.6B

"to" before an infinitive or noun

A neuron in layer 21 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 "to" when it comes right after a verb or noun and leads into the next part of a sentence, like "what we believe to be" or "put to death". Turning it up did not produce a clear theme; the text just drifted.

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

    the attributes of an educated person. The UMass Dartmouth Commitment to Student Learning, which is the foundation of the General Education Program, captures what we believe to be the essence of the UMD educated person.⏎UMass Dartmouth Commitment

  2. #2 peak

    "The median response indicated a 23 percent increase in productivity in the future atmosphere," said ORNL's Rich Norby, lead author of the paper to be published Dec. 13 in the Proceedings of the National Academy of Sciences

  3. #3 peak

    by a method unchanged since the age of the Moorish invaders _ swamping them under a flood of water from the local canal.⏎But those days may be coming to an end. Martorell is one of a growing number of Spanish farmers who

  4. #4 peak

    preventable cause of harm. To improve knowledge on anticoagulant safety, the United States (US) Office of Disease Prevention and Health Promotion issued a call to generate more evidence on anticoagulant drug interactions.⏎Potential drug-drug interactions

  5. #5 peak

    domino reaction across connected devices. In no time at all, data infrastructures can be destroyed, making the IoT one of the biggest data vulnerabilities in the years to come.⏎2. Mobile malware⏎While malicious apps

  6. #6 peak

    13th century Christian Europe willingly looked upon him as an international judge.⏎He was renowned for his charity. “The peace and blessings of the realm come to us through the poor,” he would say. Beggars were fed from his

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: When the match get get meet at a match, when the match meet match meet meet meet meet meet match meet meet meet meet meet match meet meet meet meet match meet meet meet meet match meet meet meet meet match meet meet meet meet meet match

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

    .S. history. The attack on the Greenwood section of Tulsa, home to a prosperous African American community in the segregated city, left over 35 blocks burned to the ground and over 10,000 people — most of them

  2. #2 peak

    their cows to be fed. In the back of everyone’s mind is the fact that the common is going to be ruined because the grass is going to be eaten to depletion. Still, everyone wants to get grass for their cows. No one thinks

  3. #3 peak

    space will likely always grapple with.⏎In May 2020, Rhys’ childhood home in Dominica's capital, Roseau, was demolished to accommodate commercial construction, an act that Ramlochan found “instructive”

  4. #4 peak

    towns of Hardwick and Petersham before becoming its own separate town in 1801. The town was abandoned and leveled in 1938 to make way for the Quabbin Reservoir. Much of the town, including

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. L18:1001 excites
  2. L18:2760 excites 'to' after a verb of sharing or using
  3. L17:1297 excites
  4. L19:858 suppresses
  5. L6:2000 suppresses
  6. L9:1890 excites
  7. L16:1168 excites 'to' after verbs of change
  8. L4:1035 suppresses 'for' after verbs like 'fall' or 'click'
  9. L19:1769 excites
  10. L19:2156 suppresses
  11. L20:810 suppresses
  12. L20:2814 suppresses
  13. L19:2535 excites
  14. L12:2212 weakly suppresses 'to' or 'for' opening a purpose phrase
  15. L20:2262 weakly excites
  16. L2:24 weakly suppresses
  17. L0:318 weakly excites
  18. L7:1584 weakly suppresses
  19. L0:2193 weakly excites
  20. L20:1978 weakly suppresses
  21. L5:2251 weakly suppresses
  22. L20:1623 weakly excites
  23. L16:1939 weakly excites line break mid-sentence in wrapped text
  24. L1:1900 weakly excites

Feeds later layers

  1. L22:1166 suppresses 'to' after words needing an infinitive
  2. L24:2854 suppresses
  3. L22:1806 suppresses "to" introducing a purpose clause
  4. L24:2768 suppresses
  5. L22:266 excites
  6. L22:627 suppresses
  7. L23:1169 suppresses
  8. L22:228 weakly excites
  9. L23:1115 weakly suppresses
  10. L22:200 weakly suppresses
  11. L23:1378 weakly suppresses
  12. L22:2869 weakly suppresses
  13. L24:265 weakly suppresses
  14. L24:1072 weakly suppresses
  15. L23:566 weakly suppresses
  16. L25:364 weakly suppresses
  17. L26:296 weakly suppresses
  18. L24:442 weakly suppresses
  19. L23:887 weakly suppresses
  20. L26:189 weakly excites
  21. L22:603 weakly suppresses
  22. L24:1325 weakly suppresses
  23. L25:1197 weakly suppresses
  24. L24:208 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 -28.94 and -27.951 values between -24.98 and -23.991 values between -16.08 and -15.091 values between -15.09 and -14.103 values between -13.12 and -12.131 values between -12.13 and -11.141 values between -9.16 and -8.172 values between -7.18 and -6.195 values between -6.19 and -5.213 values between -5.21 and -4.228 values between -4.22 and -3.236 values between -3.23 and -2.2414 values between -2.24 and -1.25317 values between -1.25 and -0.263447 values between -0.26 and 0.7373 values between 0.73 and 1.7216 values between 1.72 and 2.712 values between 2.71 and 3.703 values between 3.70 and 4.681 values between 4.68 and 5.671 values between 6.66 and 7.651 values between 9.63 and 10.62offnegativepositiveoften

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. Swipe
  2. 为
  3. 为进一步
  4. 是如何
  5. with
  6. cribes
  7. 为目的
  8. 起来了
  9. whose
  10. stdbool

Suppresses

  1. 避免
  2. avoid
  3. 防止
  4. 适应
  5. 期
  6. 达到
  7. avoid
  8. 逃避
  9. avoided
  10. achieve