Neuron Explorer

Qwen3-0.6B

dates and timestamps, party letters negative

A neuron in layer 11 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on the spot in a date or time stamp, such as the comma after 'September 11' or the space in 'Page 1 of 1'. It goes strongly negative on the party letter in a politician's name, like the 'D' in 'Harry Reid (D-NV)'. Turning it up or down gave rambling text that did not mention dates, so the steering does not confirm the pattern.

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

    field from August 2 – August 20.<|endoftext|>Genetics as it applies to evolution, molecular biology, and medical aspects.⏎10 posts • Page 1 of 1⏎Hope somebody can help, I've got an A level

  2. #2 peak

    // in Elliot Hildner / University of Bremen⏎The main crack<|endoftext|>analysis of algoritms⏎robert.kern at gmail.com⏎Fri Sep 10 00:10:22 CEST 20

  3. #3 peak

    New Bovine Footbath⏎Press release date: September 11, 2013⏎Plastruct Canada Inc. has broadened their

  4. #4 peak

    and PollinatorsA student-built pollinator garden is one of Muhlenberg College’s newest sustainability initiatives.⏎By: Grace Oddo '22 Wednesday, June 30, 2021 02:03

  5. #5 peak

    .codeguru.com/algorithms/factorial.shtml I don't happen to think that the code there is easily understandable though...⏎January 10th, 2004 01:32 AM⏎Tim_axe

  6. #6 peak

    are at right angles<|endoftext|>The latest news from academia, regulators⏎research labs and other things of interest⏎Posted: September 29, 2006⏎Magnetic nanobridge brings two worlds together⏎

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 need to find my own to be a student, and for my future to be in a place where I can use myself, and for my identity to be in a place where I can use myself, and for my future to be

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

    week, Senator Harry Reid (D-NV) invited more than 30 teams of researchers from the physical and

  2. #2 peak

    On the same day, the Consumer Financial Protection Bureau (CFPB) released a study reviewing financial education research priorities for America’s youth.⏎Congressmen Bill Foster (D-IL) and French Hill (R-AR) both introduced H. Res

  3. #3 peak

    , Bible is scheduled to testify at a Capitol Hill hearing before the Senate Agriculture Committee, which is considering the carbon-farming legislation. Introduced by senators Mike Braun (R-Indiana) and Debbie Stabenow (D-Michigan), the bill

  4. #4 peak

    Hill exhibition, we’re coming across new questions as fast as we can answer the original ones. There’s a long way to go yet.⏎Senator Bill Nelson (D-FL), in his keynote address, highlighted the crucial importance of providing funds to

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. L10:1491 weakly suppresses
  2. L5:1684 weakly suppresses
  3. L7:933 weakly excites
  4. L9:1067 weakly suppresses
  5. L7:1430 weakly suppresses
  6. L3:350 weakly suppresses
  7. L2:75 weakly suppresses
  8. L8:1878 weakly excites
  9. L4:993 weakly suppresses
  10. L9:1230 weakly suppresses
  11. L10:1148 weakly excites
  12. L6:591 weakly excites number and sequence words in math text
  13. L8:1614 weakly excites
  14. L2:1634 weakly suppresses
  15. L7:1973 weakly suppresses
  16. L2:2438 weakly excites
  17. L7:1202 weakly suppresses
  18. L0:66 weakly suppresses
  19. L0:2501 weakly excites
  20. L6:2178 weakly excites
  21. L8:1924 weakly suppresses
  22. L2:1043 weakly suppresses
  23. L10:450 weakly suppresses
  24. L3:2159 weakly excites

Feeds later layers

  1. L15:1912 weakly suppresses
  2. L14:2554 weakly suppresses
  3. L12:1858 weakly excites
  4. L15:1640 weakly excites
  5. L13:1329 weakly suppresses
  6. L12:1888 weakly suppresses
  7. L15:2198 weakly suppresses
  8. L14:2830 weakly suppresses
  9. L14:699 weakly excites
  10. L15:822 weakly suppresses
  11. L17:1365 weakly suppresses
  12. L13:1631 weakly suppresses
  13. L15:396 weakly suppresses
  14. L21:2070 weakly excites
  15. L17:1957 weakly suppresses
  16. L15:982 weakly excites
  17. L16:467 weakly excites
  18. L13:363 weakly excites
  19. L12:2450 weakly excites
  20. L18:1665 weakly suppresses
  21. L12:1182 weakly excites
  22. L13:323 weakly suppresses
  23. L19:700 weakly suppresses
  24. L23:590 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.

3 values between -0.78 and -0.741 values between -0.74 and -0.692 values between -0.69 and -0.651 values between -0.65 and -0.613 values between -0.56 and -0.525 values between -0.52 and -0.486 values between -0.48 and -0.438 values between -0.43 and -0.3913 values between -0.39 and -0.3528 values between -0.35 and -0.3040 values between -0.30 and -0.2678 values between -0.26 and -0.22108 values between -0.22 and -0.17164 values between -0.17 and -0.13265 values between -0.13 and -0.09340 values between -0.09 and -0.05409 values between -0.05 and -0.00547 values between -0.00 and 0.04460 values between 0.04 and 0.09422 values between 0.09 and 0.13316 values between 0.13 and 0.17220 values between 0.17 and 0.21163 values between 0.21 and 0.2698 values between 0.26 and 0.3068 values between 0.30 and 0.3459 values between 0.34 and 0.3938 values between 0.39 and 0.4319 values between 0.43 and 0.4712 values between 0.47 and 0.522 values between 0.52 and 0.562 values between 0.56 and 0.602 values between 0.60 and 0.651 values between 0.65 and 0.691 values between 0.69 and 0.731 values between 0.73 and 0.781 values between 0.78 and 0.822 values between 0.91 and 0.95offnegativepositiveoften

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. ";⏎⏎⏎
  2. ENTIAL
  3. usp
  4. '};⏎
  5. QPainter
  6. ";}⏎
  7. ();}⏎
  8. Bureau
  9. ";}
  10. "]];⏎

Suppresses

  1. owi
  2. mine
  3. isateur
  4. 越大
  5. ously
  6. [url
  7. parameters
  8. former
  9. owy
  10. sted