Neuron Explorer

Qwen3-0.6B

modifier words before their head word

A neuron in layer 14 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on a word that qualifies the word coming right after it, such as 'strongly' in 'strongly disagree' or 'very' in 'very high'. It is a loose pattern, and turning it up gave rambling or repetitive text rather than anything on theme.

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

    classroom environment to avoid having students draw comparisons between the characters in the film and their classmates.⏎Before Watching: Consider whether you strongly agree, agree, disagree, or strongly disagree with this statement: Sometimes we have to change who we truly are

  2. #2 peak

    board, each with one of the following statements written in large letters: “Strongly Agree,” “Agree Somewhat,” “Disagree Somewhat,” “Strongly Disagree”⏎Activities / Procedures:⏎1. WARM-UP/DO

  3. #3 peak

    the tau (τ−⏎) and the tau neutrino (ντ⏎). Each type of neutrino is associated with an antimatter particle, called an antineutrino⏎, which also has neutral electric charge and 1/2

  4. #4 peak

    food with either high or very high salt levels, which are comparable to a human diet, for eight weeks.

  5. #5 peak

    than a split second decision to specify materials with recycled or recyclable content. A truly sustainable strategy tries to understand the impact of building, or of not building, and to address the environmental, ecological, cultural, economic, and financial

  6. #6 peak

    states:⏎|Bit 1||Bit 0||Data Size|⏎The memory system must be able to handle word, half word and byte writes.⏎Memory systems supporting only word writes will have severe difficulties supporting

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 question, to answer? and the two parties? to the question, is the sentence in the the... and... and then, in the first paragraph, the author is? A) the the next the is

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

    .⏎Sacrificial Anodic Coatings⏎Anodic coatings for the protection of iron and steel substrates are almost totally limited to zinc and aluminium coatings or their alloys. Where coatings anodic to the substrate are applied, the corrosion protection is referred

  2. #2 peak

    think this problem is relevant for Professor Finnegan's class at WSU.⏎What textbook is this problem found in?⏎Our data indicates that this problem or a close variation was asked in Chemistry: An Atoms First Approach - Zumdahl Atoms

  3. #3 peak

    (FANC), as well as the p53 signaling pathway. B. In response to intrastrand crosslinks (crosslink), ATR and its interacting partner ATRIP become activated and phosphorylated leading to the induction of several downstream pathways

  4. #4 peak

    contain chlorine. The EPA allows a certain measurement of chlorine to be in the water and still be deemed “safe”.⏎Chlorine or Chloramines (mix of Ammonia and Chlorine that large cities use) is present as a disin

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. L11:3040 excites
  2. L13:1762 excites
  3. L12:1493 excites
  4. L12:2861 suppresses
  5. L13:2503 excites
  6. L9:946 suppresses
  7. L12:1897 suppresses
  8. L10:3054 suppresses
  9. L13:2563 excites
  10. L8:2956 excites
  11. L13:2902 suppresses
  12. L12:2221 excites
  13. L13:2435 suppresses
  14. L12:241 excites
  15. L11:1420 suppresses
  16. L11:1035 excites
  17. L13:1666 suppresses
  18. L7:753 suppresses
  19. L9:1436 excites
  20. L9:750 weakly excites
  21. L13:71 weakly suppresses
  22. L11:1770 weakly excites
  23. L9:3065 weakly excites
  24. L11:1623 weakly excites

Feeds later layers

  1. L16:2789 excites
  2. L15:1063 suppresses
  3. L15:1380 excites
  4. L17:1679 excites
  5. L15:3069 excites
  6. L17:662 excites
  7. L18:1373 excites
  8. L15:1855 excites
  9. L15:1498 weakly excites
  10. L18:1314 weakly suppresses
  11. L17:1133 weakly excites
  12. L16:561 weakly excites
  13. L15:187 weakly excites
  14. L15:1928 weakly excites
  15. L15:1208 weakly suppresses
  16. L17:3039 weakly suppresses
  17. L18:944 weakly suppresses
  18. L21:2528 weakly suppresses
  19. L19:1410 weakly excites
  20. L15:1746 weakly excites
  21. L15:1719 weakly excites
  22. L16:2086 weakly excites
  23. L15:1494 weakly suppresses
  24. L20:1157 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.

27 values between -1.06 and -0.72603 values between -0.72 and -0.382815 values between -0.38 and -0.03407 values between -0.03 and 0.3128 values between 0.31 and 0.6510 values between 0.65 and 0.996 values between 0.99 and 1.336 values between 1.33 and 1.671 values between 1.67 and 2.012 values between 2.01 and 2.361 values between 4.40 and 4.741 values between 9.52 and 9.861 values between 12.25 and 12.59offnegativepositiveoften

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. azers
  2. 删除成功
  3. 日讯
  4. 屋里
  5. 一个重要
  6. aires
  7. orphism
  8. 这几个
  9. 免责声明
  10. OperationException

Suppresses

  1. zas
  2. nees
  3. olum
  4. adia
  5. remains
  6. CTX
  7. ILE
  8. ��
  9. footing
  10. nee