Neuron Explorer

Qwen3-0.6B

opening bracket before a citation

A neuron in layer 3 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 an opening bracket that comes right after a quote or a claim and starts a citation, such as "(Morton, 1999)". Its negative side is on the name "Judas", which is a different pattern. Turning it up did not give a clear change in the writing.

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

    exists a “relation between the perceiving mind and the physical world. However, the real qualities of objects are not defined by color, warmth, and taste” (Morton, 1999 p.47). Conversely, Desc

  2. #2 peak

    et al., 2011). Paper wasps are likely to influence many other species in terrestrial ecosystems because they are voracious predators of invertebrates (Ward & Ramón-Laca, 2014). Four species

  3. #3 peak

    decline within many parts of their range (Rabinowitz & Winter, 2006).⏎Carnivore populations are critically important to maintaining healthy ecosystems (Terborgh et al., 1999; Terborgh et

  4. #4 peak

    other white men at other schools did the same. That is why we now have so many Navajo families like our own with the last name of Begay” (22). Bruchac is skilled in incorporating information into his narrative that inspires

  5. #5 peak

    poor, the present results agree with Sheriff’s criticism regarding teaching ethics, once ethics “is an inherent quality learn as part of a development process of a personality” (4). It was clear that the children knew about the prevalent types of morality in

  6. #6 peak

    hou et al. 2004; Grimm et al. 2008). Consequently, growth in urban development is a major cause of extinction (Czech & Krausman 1997), with the native species

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: I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately. I have been thinking about lately

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

    :1-2 shows that the place where Jesus was to be betrayed was a place that He went often. His disciples also went to the garden often. Judas, one of His disciples, was not just some mere acquaintance to Christ. He

  2. #2 peak

    the events. In Matthew 26:25, Judas asked Jesus if it was he who would betray Him. He had already sealed the

  3. #3 peak

    it to name their children because of its negative connotation. John 13:18-30 records the events leading up to the betrayal by Judas. Many of the details leave some

  4. #4 peak

    13 is not the first time one reads of Judas and his dishonesty. In John 12:1-8, we read that Judas was upset with Mary, the sister of Lazarus, because she was using costly

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. L2:53 weakly excites
  2. L2:111 weakly suppresses
  3. L2:2693 weakly suppresses
  4. L2:1321 weakly suppresses
  5. L1:515 weakly suppresses 'about' in 'it's not about' phrasing
  6. L1:1420 weakly excites
  7. L1:1766 weakly suppresses
  8. L0:1204 weakly suppresses
  9. L2:1702 weakly suppresses
  10. L0:1610 weakly suppresses
  11. L0:2050 weakly suppresses
  12. L1:1788 weakly excites
  13. L2:1292 weakly suppresses
  14. L1:2638 weakly excites
  15. L1:2089 weakly excites
  16. L1:2697 weakly suppresses
  17. L1:1226 weakly excites
  18. L1:1476 weakly suppresses
  19. L2:2881 weakly excites
  20. L2:845 weakly excites
  21. L1:1687 weakly excites
  22. L2:1138 weakly suppresses
  23. L1:2737 weakly suppresses
  24. L2:1178 weakly suppresses

Feeds later layers

  1. L9:159 suppresses
  2. L6:1181 excites
  3. L8:2562 weakly excites
  4. L8:248 weakly suppresses
  5. L10:2398 weakly excites
  6. L27:1899 weakly suppresses
  7. L25:1133 weakly suppresses
  8. L21:2609 weakly suppresses
  9. L8:2893 weakly suppresses
  10. L5:209 weakly excites
  11. L4:276 weakly excites
  12. L4:80 weakly excites
  13. L11:1874 weakly suppresses
  14. L4:2860 weakly suppresses
  15. L4:1453 weakly excites
  16. L19:1118 weakly suppresses
  17. L9:2544 weakly excites
  18. L13:1344 weakly excites
  19. L24:2498 weakly excites unclear: word before an event or name
  20. L13:480 weakly suppresses
  21. L22:2369 weakly suppresses
  22. L19:3003 weakly suppresses
  23. L14:1571 weakly excites
  24. L7:1730 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 -5.04 and -4.801 values between -4.80 and -4.561 values between -3.37 and -3.141 values between -1.24 and -1.003 values between -1.00 and -0.767 values between -0.76 and -0.5342 values between -0.53 and -0.29872 values between -0.29 and -0.052631 values between -0.05 and 0.19324 values between 0.19 and 0.4210 values between 0.42 and 0.662 values between 0.66 and 0.901 values between 0.90 and 1.142 values between 1.14 and 1.372 values between 1.37 and 1.612 values between 2.56 and 2.801 values between 2.80 and 3.031 values between 3.03 and 3.271 values between 3.51 and 3.751 values between 3.75 and 3.981 values between 3.98 and 4.221 values between 4.22 and 4.46offnegativepositiveoften

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. lass
  2. ted
  3. sonian
  4. antz
  5. orte
  6. entes
  7. ter
  8. endez
  9. ).⏎
  10. ).

Suppresses

  1. 其它问题
  2. itlement
  3. Mods
  4. fileName
  5. republic
  6. imeline
  7. itution
  8. uracy
  9. idelity
  10. owość