Neuron Explorer

Qwen3-0.6B

the word 'at' in place phrases

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 lights up on the small word 'at' when it starts a phrase about a place, time or level, such as 'at the end' or 'at all levels'. When it was turned down, the model got stuck repeating 'at least' over and over.

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

    . If you want the list double-spaced, highlight all the text and set the paragraph to be double-spaced. Do not simply hit the Enter button twice at the end of each line, it will become an editing nightmare and when the sorting

  2. #2 peak

    to spread renewable energy technologies to enhance energy supply in urban areas. We particularly support conceptual and analytical awareness and understanding on gender, poverty and urban energy so that authorities at all levels, civil society and private sector partners are aware and understand the linkages

  3. #3 peak

    options so you don’t get bored with your workouts. When designing your fitness plan, consider adding some stretching and flexibility training. Simply add a few minutes of stretching at the end of your cardio and weight training workouts.⏎Cardio, or aerobic,

  4. #4 peak

    .”)⏎He likes to eat popcorn when he goes to see a movie.⏎To play in the snow is one of life’s greatest pleasures.⏎(The infinitive appears at the beginning of the sentences. This sounds very formal.)⏎He went downstairs to get

  5. #5 peak

    code, x would never be zero, that wasn't the issue. The fact is that it is pointless even having the x if you set total to 0 at the start because total will never increase, and that's even if you ignore that

  6. #6 peak

    exposes them to risks. We live in a time of rapid technological and social change, and a better, safer internet for children is high on the agenda for policymakers at both the national and European level.⏎In 2013, the European

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: at least alt alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts alts

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

    functions are generaelly a subset of hash functions that fulfil at least two extra criteria:⏎In order to fulfil these criteria (or at least, as a by-product of needing to fulfil these criteria), secure hash functions

  2. #2 peak

    or disorder is the genotype with two recessive alleles. So I guess in order for the normal genotype, or what is considered normal, a person has to have at least

  3. #3 peak

    across disparate operating systems.⏎Kubernetes clusters are actually a set of nodes categorized into master nodes and worker nodes. Every cluster needs to have at least one of each. While the master node is meant to control the state of

  4. #4 peak

    of dog food, but all of them must meet certain requirements to be considered “real” dog food. For example, the food must have a meat content of at least 70% and it must not contain any artificial colors or flavors.

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. L20:2348 excites "at least" and minimum-quantity phrases
  2. L19:1171 excites
  3. L22:160 suppresses
  4. L21:1339 excites
  5. L22:1904 excites ages right before a number
  6. L23:570 excites
  7. L19:2912 excites
  8. L20:1281 suppresses
  9. L23:253 excites
  10. L20:833 excites
  11. L23:2966 suppresses
  12. L23:476 excites
  13. L21:140 excites
  14. L17:2409 suppresses
  15. L22:825 suppresses 'for' before a span of time
  16. L21:2498 excites
  17. L22:2713 excites
  18. L21:2859 excites
  19. L20:2184 excites
  20. L21:1647 weakly excites
  21. L8:2858 weakly excites
  22. L22:2720 weakly excites
  23. L16:513 weakly suppresses
  24. L22:2255 weakly suppresses

Feeds later layers

  1. L27:599 strongly suppresses
  2. L26:835 strongly suppresses
  3. L25:501 strongly excites
  4. L26:840 excites
  5. L26:1210 suppresses
  6. L26:2537 excites
  7. L26:1800 suppresses
  8. L27:1119 weakly suppresses
  9. L27:675 weakly suppresses
  10. L26:2330 weakly suppresses
  11. L26:1876 weakly suppresses
  12. L27:1315 weakly suppresses
  13. L25:582 weakly excites
  14. L27:144 weakly suppresses
  15. L26:207 weakly suppresses
  16. L26:992 weakly excites
  17. L25:3014 weakly excites
  18. L25:1145 weakly suppresses
  19. L25:2592 weakly excites
  20. L27:2999 weakly suppresses
  21. L27:1918 weakly suppresses
  22. L26:789 weakly excites
  23. L25:509 weakly suppresses
  24. L25:562 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.

2 values between -35.84 and -34.351 values between -31.35 and -29.851 values between -26.85 and -25.353 values between -25.35 and -23.851 values between -17.86 and -16.367 values between -16.36 and -14.864 values between -14.86 and -13.367 values between -13.36 and -11.865 values between -11.86 and -10.368 values between -10.36 and -8.8721 values between -8.87 and -7.3728 values between -7.37 and -5.8738 values between -5.87 and -4.37105 values between -4.37 and -2.87185 values between -2.87 and -1.372333 values between -1.37 and 0.13801 values between 0.13 and 1.63183 values between 1.63 and 3.1386 values between 3.13 and 4.6350 values between 4.63 and 6.1215 values between 6.12 and 7.629 values between 7.62 and 9.127 values between 9.12 and 10.622 values between 10.62 and 12.121 values between 12.12 and 13.624 values between 13.62 and 15.121 values between 22.61 and 24.11offnegativepositiveoften

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. juana
  2. onian
  3. udad
  4. headed
  5. onden
  6. DialogContent
  7. onomies
  8. INCT
  9. CUS
  10. ąż

Suppresses

  1. at
  2. At
  3. At
  4. _at
  5. →at
  6. at
  7. -at
  8. .at
  9. (at
  10. [at