Neuron Explorer

Qwen3-0.6B

the blank slot before a year

A neuron in layer 18 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 spot just before a year number, such as the gap in 'in [blank]2015' or 'the year [blank]2008'. It is the point where the model is about to write a year. Turning it down made the model keep writing years, and the negative side is on numbers and words in unit or measurement contexts.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

    be aware of your weaknesses and thus exploit them more effectively than an outside agent might be able to.⏎This chart shows the type of people that committed cyberattacks in 2015.<|endoftext|>NPR’s Fresh Air book review mentions Georgia Ch

  2. #2 peak

    , multi-scale nested watersheds were delineated and the relationships between two representative water quality indexes and agricultural land use intensity were assessed and quantified for the year 2000 using multi-scale regression analysis. The results show that the log

  3. #3 peak

    district come from a variety of sources. For income taxes paid, we used the most recent dataset released by the IRS, which breaks down tax information from the year 2007 by zip code. We then link zip codes to congressional districts

  4. #4 peak

    grown, its headquarters have expanded into the surrounding area. In 1917, an administration building was built on the block east of the temple and in 1972, the twenty-eight story LDS Church Office Building, which was

  5. #5 peak

    . Studies show that though only 31.4 per cent of the adult populace had teeth rotting in the year 1994; the year 2008 saw this percentage increase to a shocking 39.9

  6. #6 peak

    ry disease characteristics, and the clinical parameters.⏎The investigators scoured the Pubmed, EMBASE, and CINAHL databases from their inception to February of 2018 to identify articles on Fabry disease and white matter lesions examined

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 a 30000000000000000000000000000000000000000000

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

    phenomena, directly and indirectly, constitute active transport of matter due to an applied electric current. The transport is measured in units of chemical flux, commonly µmol/cm2h. Iontophoresis has experimental, therapeutic and diagnostic applications.⏎I

  2. #2 peak

    Glenn Fowler, M.D., M.P.H., of the Makerere University–Johns Hopkins University Research Collaboration in Kampala, Uganda, the study team expects results in five to six years.⏎The HIV-infected women eligible to participate in PROM

  3. #3 peak

    are released. The eggs are semi-buoyant, causing them to be dispersed widely by the water current. The eggs develop rapidly, hatching into larvae within 24 to 48 hours after fertilisation (8).<|endoftext|>Peter

  4. #4 peak

    is a communicable disease, which is caused by a recently revealed coronavirus. Those who are infected with the virus will experience mild to modest respiratory sickness. They can recover from the illness without the need for special treatment. The older population is more vulnerable

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. L17:2964 strongly suppresses
  2. L17:2750 strongly suppresses
  3. L17:2760 strongly excites
  4. L17:2544 strongly suppresses
  5. L12:2593 strongly excites
  6. L12:2691 strongly suppresses
  7. L14:2201 excites
  8. L16:1815 excites
  9. L15:1520 excites
  10. L17:1041 excites
  11. L12:2338 excites
  12. L15:2447 suppresses
  13. L14:2939 excites
  14. L14:1088 excites
  15. L10:637 suppresses
  16. L14:1778 suppresses
  17. L15:1825 excites
  18. L10:986 excites
  19. L11:1352 excites
  20. L11:2530 excites
  21. L16:1050 suppresses
  22. L16:1668 suppresses
  23. L16:1930 excites
  24. L15:2851 suppresses

Feeds later layers

  1. L22:1074 strongly suppresses
  2. L25:565 suppresses
  3. L21:741 suppresses
  4. L19:245 excites
  5. L20:1680 suppresses
  6. L20:2472 suppresses
  7. L21:848 suppresses
  8. L20:2347 suppresses
  9. L22:2371 suppresses
  10. L22:2932 suppresses
  11. L19:2242 suppresses
  12. L20:2792 suppresses
  13. L26:284 suppresses
  14. L24:906 suppresses
  15. L21:1729 suppresses
  16. L22:1157 suppresses
  17. L20:1335 suppresses
  18. L22:2445 suppresses
  19. L23:1919 suppresses
  20. L27:934 suppresses
  21. L19:1720 suppresses
  22. L25:2440 suppresses
  23. L23:1337 suppresses
  24. L19:2474 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 -10.11 and -9.391 values between -9.39 and -8.681 values between -7.25 and -6.532 values between -6.53 and -5.812 values between -5.81 and -5.102 values between -5.10 and -4.384 values between -4.38 and -3.676 values between -3.67 and -2.9518 values between -2.95 and -2.2331 values between -2.23 and -1.5280 values between -1.52 and -0.80807 values between -0.80 and -0.092768 values between -0.09 and 0.63108 values between 0.63 and 1.3515 values between 1.35 and 2.0614 values between 2.06 and 2.7810 values between 2.78 and 3.508 values between 3.50 and 4.212 values between 4.21 and 4.933 values between 4.93 and 5.647 values between 5.64 and 6.363 values between 6.36 and 7.083 values between 7.08 and 7.792 values between 7.79 and 8.511 values between 8.51 and 9.224 values between 9.22 and 9.942 values between 9.94 and 10.661 values between 10.66 and 11.371 values between 12.09 and 12.801 values between 17.82 and 18.53offnegativepositiveoften

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. ambi
  2. ButtonTitles
  3. 的手
  4. ragments
  5. enticated
  6. 幂
  7. getState
  8. 貌
  9. мот
  10. 一脚

Suppresses

  1. time
  2. 时间
  3. 时间
  4. 的时间
  5. stime
  6. 时间为
  7. etime
  8. Time
  9. _Time
  10. $time