Neuron Explorer

Qwen3-0.6B

"the" after certain verbs and prepositions

A neuron in layer 18 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 "the" in certain fixed spots, such as "like the" or "comes from the". Turning it up made the model repeat phrases about "the problem" or "the time".

Written by Claude from this neuron's activations and steering outputs. Confidence: low. 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

    blizzard shuts down the highway, Phil finds himself trapped in Punxsutawney. He wakes up the next day, only to discover that it's not the next day at all. It's Groundhog Day all over again.⏎via Wikimedia

  2. #2 peak

    Diagnosis and Treatment Interventions for Racial and Ethnic Minorities<|endoftext|>Lily of the valley, or convallaria, is not a true lily like the Easter or Asiatic lily, but is still toxic to dogs, humans and

  3. #3 peak

    be rejuvenated in any way.⏎Likewise, damage by harsh scratching cannot be reversed.⏎Scratched tiles will continue to perform<|endoftext|>The word biathlon comes from the Greek word meaning two contests. At the Winter Olympics it means joining two sports together

  4. #4 peak

    ithridatic War. As a result, Pompey had annexed Syria as a new<|endoftext|>Our Earth is one among many heavenly bodies rotating around a star, the Sun, which gives us light and heat. They all form what we call our

  5. #5 peak

    and the Israelite people. Curiously, it was the king of Egypt, and not God, or the Israelites themselves, who coined

  6. #6 peak

    today."⏎It is a question that will resonate with millions living in quarantine today – as people wake up every morning wondering if the day ahead will be any different from the 24 hours they have just endured.⏎But I have a more positive spin

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 problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of the problem of

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

    Missouri in 2015. They are the same facilities as in 2013, except that the Montrose Generating Station has dropped out of the top 10, and the Mississippi Lime Company

  2. #2 peak

    as heart disease.⏎This study looked only at women diagnosed with breast cancer early – known as stages one or two. It did not include patients with advanced forms of the illness.⏎Experts believe radiotherapy may be far more effective at killing all cancerous

  3. #3 peak

    – is rooted in a dispute over land claimed by Jews as their biblical birthright and by the Palestinians, who seek self-determination. Despite repeated attempts to end the conflict between the two countries, there is no peace settlement in sight. The issue

  4. #4 peak

    for you. Thank You.⏎When I first entered the spoken word and writing community in the mid-1990s, one of the biggest names in the Southern California scene was Michele Serros. The Oxnard-born Serros great

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. L16:792 excites "the" inside a prepositional phrase
  2. L17:30 suppresses
  3. L16:419 excites
  4. L10:1377 suppresses
  5. L9:662 suppresses
  6. L12:435 suppresses
  7. L11:1168 excites
  8. L16:1320 suppresses
  9. L4:52 excites
  10. L9:24 excites "the" before a named organisation or title
  11. L8:636 suppresses
  12. L10:120 suppresses
  13. L17:372 suppresses
  14. L11:331 excites
  15. L16:2248 suppresses
  16. L9:2198 excites
  17. L13:824 suppresses
  18. L6:443 suppresses
  19. L16:2062 excites "other" and contrast words
  20. L11:875 weakly suppresses
  21. L10:1644 weakly excites
  22. L3:344 weakly excites
  23. L15:1848 weakly suppresses
  24. L13:1234 weakly suppresses

Feeds later layers

  1. L21:843 suppresses
  2. L19:2430 suppresses
  3. L20:1197 suppresses 'this' pointing at the current document
  4. L20:1983 suppresses
  5. L19:1032 suppresses
  6. L19:2638 suppresses
  7. L20:714 weakly suppresses
  8. L19:2537 weakly suppresses
  9. L19:2958 weakly suppresses
  10. L23:1817 weakly suppresses
  11. L19:326 weakly excites
  12. L21:713 weakly excites
  13. L24:1416 weakly excites
  14. L20:47 weakly suppresses
  15. L19:1900 weakly suppresses
  16. L19:1159 weakly excites
  17. L20:337 weakly excites
  18. L27:533 weakly suppresses
  19. L22:2298 weakly excites
  20. L20:2860 weakly excites
  21. L24:704 weakly excites
  22. L23:355 weakly suppresses
  23. L19:1023 weakly suppresses
  24. L21:639 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 -14.34 and -13.732 values between -13.12 and -12.511 values between -12.51 and -11.893 values between -11.89 and -11.281 values between -10.06 and -9.441 values between -9.44 and -8.835 values between -8.83 and -8.225 values between -8.22 and -7.602 values between -7.60 and -6.9911 values between -6.99 and -6.3810 values between -6.38 and -5.7710 values between -5.77 and -5.159 values between -5.15 and -4.5410 values between -4.54 and -3.9312 values between -3.93 and -3.3213 values between -3.32 and -2.7014 values between -2.70 and -2.0921 values between -2.09 and -1.4827 values between -1.48 and -0.86226 values between -0.86 and -0.252648 values between -0.25 and 0.36792 values between 0.36 and 0.9746 values between 0.97 and 1.599 values between 1.59 and 2.204 values between 2.20 and 2.819 values between 2.81 and 3.427 values between 3.42 and 4.042 values between 4.04 and 4.652 values between 4.65 and 5.261 values between 5.26 and 5.881 values between 5.88 and 6.491 values between 7.71 and 8.331 values between 8.33 and 8.941 values between 9.55 and 10.16offnegativepositiveoften

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. ason
  2. InView
  3. ляр
  4. &o
  5. asons
  6. &a
  7. courses
  8. نظم
  9. ển
  10. одеж

Suppresses

  1. ronym
  2. 事
  3. 冒
  4. procedure
  5. piel
  6. kh
  7. 其
  8. matter
  9. suck
  10. ilde