Neuron Explorer

Qwen3-0.6B

open bracket after a value or citation

A neuron in layer 14 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 opening bracket that comes right after a number, measurement or citation, such as the '(' in '73.0°F (22.8°C)'. It switches off, and sometimes turns negative, on the word 'as' in phrases like 'referred to as' or 'described as'. Steering did not show a clear visible effect.

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

    protected by the Second Amendment. An object is “dangerous” when it is “likely to cause serious bodily harm.” Black’s Law Dictionary 451 (9th ed. 200

  2. #2 peak

    bonds with a magnetic field via a device such as a laser. The record for the hottest plasma formed on Earth measures a scorching 6 trillion degrees Celsius (10.8 trillion °F)! Definitely hotter than the temperature of the centre

  3. #3 peak

    innovative drugs have been introduced, from Pfizer's (<|endoftext|>Information identified as archived on the Web is for reference, research or recordkeeping

  4. #4 peak

    anna Climate).⏎The average temperature for the year in Mount Surprise is 73.0°F (22.8°C). The warmest month, on average, is November

  5. #5 peak

    the Internet: legal frames and technological implications’ organized by the Higher School of Economics on the 7th Internet Governance Forum (Baku, Azerbaijan, 6-9 November, 2012).⏎The emergence of the Internet is a

  6. #6 peak

    within a matter of seconds. See George C. Wilson, Visible Violence, 12 NAT’L J. 886, 887 (2003). Short of bombs, missiles, and biochemical agents, we

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: What the question, “Is the A+? Is the B+? Is the positive? What, what? What? What? What? What? I had the thought... I was thinking... I was... thinking... I was...

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

    forename – whereas “Brown” is the surname or family name.⏎Is surname your last name?⏎In the English-speaking world, a surname is commonly referred to as a last name because it is usually placed at the end of a person’s full

  2. #2 peak

    from several generations of dark, fluid basalt lava flows; the flows surround the 322--meter high stratovolcano (Jabal is translated as "mountain" in Arabic). Jabal Abyad, in the center

  3. #3 peak

    . It was first described from samples collected on Ernest H Shackleton’s expedition to the Antarctic in 1907.⏎Cyanobacteria were initially described as algae in the 18th century, before scientists realised they were bacteria.

  4. #4 peak

    occasionally, escapes occur!). These are the timber rattlesnake, the massasauga (erroneously called “pygmy rattler”), and the copperhead. All three are uncommon

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. L13:897 strongly excites
  2. L6:858 excites
  3. L4:24 excites
  4. L4:20 suppresses
  5. L8:137 suppresses
  6. L10:417 suppresses
  7. L6:2237 excites
  8. L13:287 suppresses
  9. L9:66 excites
  10. L5:1998 excites
  11. L10:272 suppresses
  12. L9:319 excites equals signs in maths and equations
  13. L6:1008 excites 'of' after a measured quantity
  14. L12:1150 suppresses
  15. L6:2091 suppresses
  16. L5:1395 excites
  17. L5:1016 excites
  18. L9:2096 excites 'the' or 'as' before a superlative role
  19. L8:749 suppresses
  20. L5:589 excites
  21. L8:2127 excites
  22. L3:76 suppresses
  23. L4:116 suppresses
  24. L9:965 excites

Feeds later layers

  1. L17:882 strongly excites
  2. L16:2038 suppresses
  3. L16:607 suppresses
  4. L17:264 excites
  5. L19:2450 suppresses
  6. L16:3046 suppresses
  7. L19:1613 suppresses "is" after a sentence-opening "It"
  8. L16:466 suppresses 'is' after a long subject phrase
  9. L15:218 excites
  10. L16:1830 suppresses
  11. L18:115 suppresses
  12. L15:1223 excites
  13. L16:2039 suppresses
  14. L20:880 suppresses
  15. L19:2553 suppresses
  16. L20:1253 suppresses
  17. L17:325 suppresses
  18. L16:2444 suppresses
  19. L18:857 excites
  20. L19:255 suppresses
  21. L16:1147 suppresses
  22. L16:316 suppresses
  23. L16:1124 excites
  24. L19:272 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 -4.49 and -4.331 values between -3.82 and -3.662 values between -3.49 and -3.323 values between -3.15 and -2.992 values between -2.99 and -2.826 values between -2.82 and -2.651 values between -2.65 and -2.497 values between -2.49 and -2.322 values between -2.32 and -2.157 values between -2.15 and -1.9810 values between -1.98 and -1.825 values between -1.82 and -1.6510 values between -1.65 and -1.4811 values between -1.48 and -1.3213 values between -1.32 and -1.1516 values between -1.15 and -0.9823 values between -0.98 and -0.8118 values between -0.81 and -0.6527 values between -0.65 and -0.4835 values between -0.48 and -0.31150 values between -0.31 and -0.141057 values between -0.14 and 0.021464 values between 0.02 and 0.19787 values between 0.19 and 0.36207 values between 0.36 and 0.5330 values between 0.53 and 0.696 values between 0.69 and 0.862 values between 0.86 and 1.033 values between 1.03 and 1.191 values between 1.36 and 1.531 values between 2.03 and 2.20offnegativepositiveoften

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. يق
  2. strup
  3. хват
  4. sucker
  5. .fa
  6. ranges
  7. icides
  8. 有两个
  9. ncpy
  10. PermissionsResult

Suppresses

  1. 一场
  2. 什么样
  3. 做什么
  4. 何种
  5. 何处
  6. 什麽
  7. 哪家
  8. 些什么
  9. 干什么
  10. yours