Neuron Explorer

Qwen3-0.6B

plural '(s)' in parentheses

A neuron in layer 16 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 lights up on the bracketed '(s)' that writers add to words like 'agent(s)' or 'pin(s)' when they are not sure whether something is singular or plural. When it was turned up, the model's writing drifted into odd, repetitive text that had nothing to do with the topic.

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

    nine were found to be “obstructive.”<|endoftext|>Rural earthen roads impact assessment in Phewa watershed, Western region, Nepal⏎© The Author(s). 2016⏎Received: 23 November 20

  2. #2 peak

    to stimulation or verbal commands. It includes administration of sedative and/or analgesic agent(s) by a route other than IV; (PO, PR, Intranasal

  3. #3 peak

    17.0.0/17⏎- We need a network with <|endoftext|>|Classification||Electronic musical instrument, synthesizer|⏎|Inventor(s)||Suzuki Musical Instrument Corporation|⏎|Tronichord, autoh

  4. #4 peak

    ’s airway, protective reflexes and the ability to respond to stimulation or verbal commands. It includes intravenous administration of sedative and/or analgesic agent(s) and appropriate monitoring.⏎Local Anesthesia: The loss of pain sensation over a

  5. #5 peak

    to 3.66 m) in width, and 8 feet to 12 feet (2.44 m to 3.66 m) in height. These measurements are common wall and roof dimensions in building design.

  6. #6 peak

    . The drops must be warmed to body temperature by putting the bottle in your hand for a few seconds before use. Before administering the drops, washing the affected ear(s) and drying them with a clean towel is necessary. To administer

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 same problem, and the problem of the same problem, and the problem of the same problem, and the problem of the same problem, the same problem, all the problem of the same problem, all the problem of the

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

    ter, Taslima and Simonovic, Slobodan P., "A General Overview of Multi-objective Multiple-participant Decision Making for Flood Management" (2002). Water Resources

  2. #2 peak

    the Australian population (Australian Journal of General Practice (2021) but many of us don’t really understand what this diagnosis

  3. #3 peak

    1. Which means they proved that a(30) ≤ 16.Share:<|endoftext|>Communication for Empowerment⏎Communication for Empowerment (C4E) is one of the practical approaches developed by OGC to address

  4. #4 peak

    groundwater for hydro electric power generation⏎Bardsley, W.E. (1995). Using groundwater for hydro electric power generation. Journal of Hydrology (NZ), 34(1), 1-14.⏎Permanent Research

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. L10:107 strongly suppresses unclear: spots where text was removed
  2. L7:701 excites
  3. L15:1374 excites
  4. L11:1579 suppresses
  5. L3:1388 excites
  6. L10:142 suppresses opening parenthesis mid-sentence
  7. L15:650 excites
  8. L1:102 excites
  9. L9:632 suppresses
  10. L4:1572 excites
  11. L3:10 excites
  12. L10:43 excites "terms" in "in terms of"
  13. L5:2525 suppresses
  14. L12:50 suppresses
  15. L15:2795 suppresses
  16. L15:608 suppresses
  17. L9:982 suppresses
  18. L6:293 suppresses
  19. L13:2197 suppresses
  20. L6:468 excites
  21. L8:62 suppresses
  22. L8:79 suppresses
  23. L9:2249 suppresses
  24. L11:2783 excites

Feeds later layers

  1. L19:471 strongly excites
  2. L18:678 strongly excites
  3. L20:131 excites
  4. L20:306 excites
  5. L19:353 excites
  6. L17:776 excites
  7. L22:765 excites
  8. L17:70 excites
  9. L23:191 excites
  10. L20:597 excites
  11. L22:161 excites
  12. L24:508 excites
  13. L20:892 suppresses
  14. L21:189 excites document break token mid-sentence
  15. L25:17 excites
  16. L19:278 excites
  17. L26:326 excites
  18. L19:204 excites
  19. L24:39 excites
  20. L24:392 excites
  21. L26:836 weakly excites
  22. L23:517 weakly excites
  23. L23:290 weakly excites
  24. L18:2164 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.

2 values between -2.57 and -2.311 values between -2.06 and -1.801 values between -1.80 and -1.543 values between -1.54 and -1.2832 values between -1.28 and -1.02193 values between -1.02 and -0.77497 values between -0.77 and -0.51688 values between -0.51 and -0.251464 values between -0.25 and 0.01868 values between 0.01 and 0.2791 values between 0.27 and 0.5318 values between 0.53 and 0.7913 values between 0.79 and 1.044 values between 1.04 and 1.306 values between 1.30 and 1.562 values between 1.56 and 1.822 values between 1.82 and 2.083 values between 2.08 and 2.343 values between 2.34 and 2.594 values between 2.59 and 2.852 values between 2.85 and 3.111 values between 3.11 and 3.371 values between 3.63 and 3.881 values between 3.88 and 4.142 values between 4.14 and 4.401 values between 4.92 and 5.181 values between 5.95 and 6.211 values between 6.73 and 6.981 values between 6.98 and 7.241 values between 7.24 and 7.501 values between 7.50 and 7.76offnegativepositiveoften

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. oplevel
  2. condolences
  3. Equality
  4. ichten
  5. anken
  6. relationship
  7. ้ว
  8. 값을
  9. ProcessEvent
  10. vows

Suppresses

  1. Cop
  2. !)
  3. !)⏎⏎
  4. ège
  5. Sophie
  6. Ģ
  7. 预
  8. yat
  9. би
  10. Mart