Neuron Explorer

Qwen3-0.6B

opening parenthesis mid-sentence

A neuron in layer 10 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 reacts to an opening bracket '(' that starts an aside or a short abbreviation, like '(USDA)' or '(for a long time, at least)'. The same bracket can push the neuron either way, and steering did not produce a clear theme.

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

    by itself. Therefore, a number of the more popular study questions include (but are not limited to) whether there are connections between colors and learning, whether

  2. #2 peak

    ght” in 1993, he became one of the first to author the word aight in published text. He’s even glossed it “(Alright)” for everyone who hadn’t caught on yet.⏎We learn from the O

  3. #3 peak

    . (This also happens to be the opinion of the FDA, public health officials, and

  4. #4 peak

    . (June 2009) (Learn how and when to remove this template

  5. #5 peak

    unctionals show that sometimes simply tethering two useful molecules together unlocks some useful activity⏎There’s been a real surge in drug discovery involving a class of molecules that (for a long time, at least) were considered at best an academic curiosity.

  6. #6 peak

    ) the right to amend that record if it is inaccurate, irrelevant, untimely, or incomplete, and (3) the right to sue the government for violations of the statute including permitting others

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 potential of a computer) or a computer) or a computer) or acomputer) or acomputer) ora computer) ora computer) ora computer) ora computer) ora computer) ora computer) |a

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

    ilohertz (kHz) passive integrated transponder (PIT) tag. Twenty subadults of each species were implanted with short-term (3 month) acoustic telemetry tags to examine dispersal patterns immediately following release. Ten

  2. #2 peak

    you to eat more produce: The government has gotten on the bandwagon. According to the U.S. Department of Agriculture’s (USDA) and Department of Health and Human Services’ (HHS) Dietary

  3. #3 peak

    and the belief that we can recover from setbacks and learn from them.⏎Singaporean students have typically scored very well in the Organisation for Economic Co-operation and Development’s (OECD) Programme for International Student Assessment (PISA) in Reading, Mathematics

  4. #4 peak

    over a period of 15 years (1996 to 2011) with different levels of harvesting intensity, including selective harvesting at low (13.0% removal of growing stock volume), medium (29.

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. L5:243 suppresses
  2. L9:339 excites
  3. L3:10 excites
  4. L8:189 suppresses
  5. L9:2300 excites
  6. L4:769 suppresses
  7. L9:2249 suppresses
  8. L6:854 excites comma after an introductory word like then or when
  9. L9:1371 excites
  10. L0:86 excites
  11. L7:288 excites
  12. L4:2293 excites
  13. L5:937 suppresses
  14. L4:2958 excites
  15. L2:353 excites
  16. L5:275 excites
  17. L9:284 suppresses
  18. L6:144 excites
  19. L3:2051 excites
  20. L4:64 excites
  21. L4:2309 excites
  22. L1:102 excites
  23. L0:1860 weakly excites
  24. L9:935 weakly suppresses

Feeds later layers

  1. L14:90 strongly suppresses
  2. L15:770 strongly suppresses
  3. L16:455 strongly suppresses
  4. L16:264 suppresses opening parenthesis in citations and asides
  5. L16:1767 suppresses
  6. L18:1365 suppresses
  7. L18:1140 suppresses
  8. L18:1247 suppresses
  9. L16:183 excites
  10. L19:1099 suppresses
  11. L18:357 suppresses
  12. L13:1721 excites
  13. L16:1253 suppresses
  14. L17:238 suppresses
  15. L11:242 suppresses
  16. L16:332 suppresses plural '(s)' in parentheses
  17. L16:2853 suppresses
  18. L19:1762 suppresses
  19. L18:721 suppresses opening parenthesis after a word
  20. L14:220 suppresses
  21. L15:277 suppresses
  22. L19:1864 suppresses
  23. L16:1494 suppresses
  24. L15:1193 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 -6.65 and -6.461 values between -5.71 and -5.521 values between -4.96 and -4.772 values between -4.21 and -4.032 values between -3.84 and -3.653 values between -3.09 and -2.901 values between -2.90 and -2.721 values between -2.72 and -2.535 values between -2.34 and -2.152 values between -2.15 and -1.974 values between -1.97 and -1.784 values between -1.78 and -1.596 values between -1.59 and -1.418 values between -1.41 and -1.223 values between -1.22 and -1.036 values between -1.03 and -0.845 values between -0.84 and -0.668 values between -0.66 and -0.4726 values between -0.47 and -0.2826 values between -0.28 and -0.10273 values between -0.10 and 0.091541 values between 0.09 and 0.281574 values between 0.28 and 0.47378 values between 0.47 and 0.6527 values between 0.65 and 0.84offnegativepositiveoften

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. ovy
  2. (Parameter
  3. `'
  4. ammen
  5. GORITHM
  6. owitz
  7. ��
  8. `.
  9. hyper
  10. cyl

Suppresses

  1. int
  2. 受
  3. indo
  4. 礼
  5. ida
  6. ()},
  7. *);⏎
  8. }",
  9. frog
  10. ”。⏎⏎