Neuron Explorer

Qwen3-0.6B

the digit 1 in years

A neuron in layer 0 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 the digit 1 when it is part of a year such as 2011 or 2021, and also to the 1 in codes like COVID-19. When it was turned up, the model's replies did not show a clear shared 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

    to the U.S. census. Without adequate paid maternity leave, mothers are frequently forced to quit stable jobs and disrupt career plans. Add to that $611 in average monthly childcare costs (U.S. Department of Labor), which is more

  2. #2 peak

    Journal of Biogeography, 48(7), 1594-1605 https://doi.org/10.1111/jbi.14097⏎

  3. #3 peak

    family studies⏎Langley, Elizabeth Dale, "Parenting Styles and Perceptions Across Christian Denominations" (2011). ETD Collection for Pace University. AAI346737

  4. #4 peak

    prices and marginal speculation in food have done the same.⏎2011 also saw the highest average oil prices since 2008, at $

  5. #5 peak

    ||Water (%)||1.57% (1,014 km2)|⏎|-||2011 estimate||2,229,641 (143rd

  6. #6 peak

    thrusts those alternate plans into our faces, which can make some people who already have self-esteem issues feel insecure about their social interactions.⏎In 2011 at the Annual Convention of the American Psychological Association, Larry D. Rosen, who

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, and I have been thinking about lately. So I think that I have been thinking about lately. I think that I have been thinking about lately. I think that I have been thinking about lately. I think

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

    10 @ 8:30 pm In General | 4 Comments⏎Across the globe, young people are paying attention to what is happening at the COP16 climate negotiations in Cancun, Mexico. They are the “climate generation”

  2. #2 peak

    some circumstances if you are carrying or showing symptoms of coronavirus (COVID-19). Face coverings should not be used as a replacement for social distancing or

  3. #3 peak

    will have come up with a workable topic for their essay and learned to focus on points in their lives, rather than the entirety of it.<|endoftext|>COVID-19 Vaccine Enters Final Test Phase⏎The COVID-19 pandemic seemingly has

  4. #4 peak

    have been given — a weevil species with their own name in the backyard.<|endoftext|>February 16, 2021⏎COVID-19 and the unexpected switch to remote learning have negatively impacted college students’ academic performance

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

None: this is the first layer.

Feeds later layers

  1. L4:1721 suppresses the gap before a decade or century
  2. L18:1361 weakly excites
  3. L9:657 weakly excites
  4. L3:1035 weakly suppresses
  5. L25:1965 weakly suppresses the digit 1 in age ranges
  6. L21:2912 weakly suppresses
  7. L16:1220 weakly suppresses
  8. L7:1001 weakly suppresses
  9. L8:457 weakly suppresses year digits in copyright and dates
  10. L23:286 weakly suppresses
  11. L10:2418 weakly suppresses
  12. L5:1090 weakly suppresses
  13. L18:824 weakly excites
  14. L26:1465 weakly suppresses
  15. L6:1683 weakly excites
  16. L19:2153 weakly suppresses
  17. L15:442 weakly suppresses
  18. L20:2620 weakly suppresses
  19. L5:2150 weakly suppresses
  20. L9:275 weakly suppresses
  21. L14:2447 weakly excites
  22. L27:2519 weakly suppresses
  23. L3:88 weakly suppresses
  24. L21:1365 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.

2 values between -1.12 and -1.053 values between -0.97 and -0.893 values between -0.89 and -0.815 values between -0.81 and -0.735 values between -0.73 and -0.665 values between -0.66 and -0.581 values between -0.58 and -0.503 values between -0.50 and -0.424 values between -0.42 and -0.3411 values between -0.34 and -0.2631 values between -0.26 and -0.1947 values between -0.19 and -0.11499 values between -0.11 and -0.032718 values between -0.03 and 0.05473 values between 0.05 and 0.1360 values between 0.13 and 0.2117 values between 0.21 and 0.282 values between 0.28 and 0.361 values between 0.36 and 0.441 values between 0.44 and 0.521 values between 0.52 and 0.601 values between 0.60 and 0.682 values between 0.75 and 0.831 values between 0.83 and 0.911 values between 0.99 and 1.071 values between 1.07 and 1.142 values between 1.14 and 1.221 values between 1.54 and 1.611 values between 1.61 and 1.692 values between 1.69 and 1.771 values between 1.77 and 1.852 values between 1.85 and 1.931 values between 1.93 and 2.00offnegativepositiveoften

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. lias
  2. 查询
  3. 霄
  4. 吾
  5. Ra
  6. ession
  7. 者
  8. throughout
  9. itta
  10. iro

Suppresses

  1. lemn
  2. pregn
  3. SplitOptions
  4. shouldBe
  5. ')['
  6. .nz
  7. ertain
  8. 在这种
  9. 多い
  10. .Lines