Neuron Explorer

Qwen3-0.6B

'the' in 'in the long run'

A neuron in layer 26 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 word 'the' in phrases like 'in the long run' or 'in the end', where 'in the' means 'eventually'. Turning it up made the model get stuck repeating words like 'medium' or 'very', which does not relate to the trigger.

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

    of your diet. Although processed food has an undeniable convenience factor and longer shelf life, they typically contain unnecessary additives and chemicals that can be detrimental to your health in the long run.⏎While not always the budget-friendly option, there are many benefits to

  2. #2 peak

    ay/grain, and finally in extreme cases you might even consider selling productive livestock. In many cases, making the hard decisions early can be the best decision in the long-run.⏎Be ready with nitrogen⏎Pastures and livestock can make compensatory

  3. #3 peak

    solar energy is environmentally friendly. It does not produce harmful emissions like some other forms of energy production do. The third reason is that solar energy is cost-effective in the long run.⏎Solar mobile power sources are a cost-effective alternative to UPS systems⏎

  4. #4 peak

    they would be more efficient and the world would be slightly more sustainable in the reduction of energy and time used by humans to create robots for a better world.

  5. #5 peak

    . This may lead to a difficulty in controlling your metabolism and a general slowing of the metabolism. The latter of these typically makes it much harder to lose weight in the end. Yoyo dieting may also put other strains on different parts of your

  6. #6 peak

    which is more resilient and cost-effective in the long term. Local governments need systems that enable them to incorporate the health of natural

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: If I really want to understand what reallyiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaiaia

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

    in their community. I love the story of Luiz.⏎Lisa Baldwin, a chemistry teacher, works with her students to fight through academic challenges.⏎By MATT RICHTEL⏎There is a widespread

  2. #2 peak

    content, color, weight, and construction all factor into what level is offered.⏎Based on ASTM International testing, products that feature l<|endoftext|>How Dick Cheney & Donald Rumsfeld Helped Richard Nixon Implement The Worst Idea Ever⏎Conor Frieders

  3. #3 peak

    The Big Picture? An Afterword⏎PortraitDyan Mazurana is senior research fellow at the Feinstein International Famine Center, Tufts University. Angela Raven-Roberts is director of research and programs at the Feinstein Famine Center, Tu

  4. #4 peak

    burg and Iredell, and several mines, some still in operation, in Stanly county in the New London/Richfield area. But Gold Hill, just a few miles from Stanly and

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. L16:107 excites
  2. L20:185 weakly excites
  3. L19:572 weakly excites
  4. L2:2028 weakly suppresses
  5. L11:2758 weakly suppresses
  6. L17:166 weakly suppresses
  7. L18:282 weakly suppresses
  8. L5:2175 weakly suppresses
  9. L10:1745 weakly suppresses
  10. L24:1223 weakly suppresses
  11. L20:164 weakly suppresses
  12. L16:305 weakly suppresses
  13. L20:1713 weakly suppresses
  14. L19:277 weakly excites
  15. L2:1866 weakly suppresses
  16. L21:2096 weakly suppresses
  17. L16:463 weakly suppresses 'for' introducing a purpose or target
  18. L16:137 weakly excites
  19. L17:253 weakly suppresses
  20. L0:41 weakly suppresses
  21. L18:137 weakly excites 'the' after 'to' in fixed phrases
  22. L10:12 weakly suppresses
  23. L22:316 weakly suppresses
  24. L10:503 weakly excites

Feeds later layers

  1. L27:1178 strongly excites
  2. L27:1745 weakly excites
  3. L27:2621 weakly suppresses
  4. L27:842 weakly excites
  5. L27:2280 weakly excites
  6. L27:1382 weakly excites
  7. L27:1221 weakly excites
  8. L27:2747 weakly suppresses
  9. L27:1165 weakly excites
  10. L27:1810 weakly excites
  11. L27:2240 weakly excites
  12. L27:2429 weakly excites
  13. L27:2638 weakly excites
  14. L27:930 weakly excites
  15. L27:1089 weakly excites
  16. L27:2885 weakly suppresses
  17. L27:1339 weakly suppresses
  18. L27:837 weakly suppresses
  19. L27:274 weakly excites document boundary, code keywords next
  20. L27:333 weakly excites
  21. L27:3054 weakly excites
  22. L27:2600 weakly excites
  23. L27:1733 weakly excites
  24. L27:1951 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 -20.36 and -19.221 values between -16.95 and -15.813 values between -14.67 and -13.542 values between -13.54 and -12.401 values between -12.40 and -11.261 values between -11.26 and -10.132 values between -10.13 and -8.997 values between -8.99 and -7.8514 values between -7.85 and -6.7116 values between -6.71 and -5.5837 values between -5.58 and -4.4461 values between -4.44 and -3.30112 values between -3.30 and -2.17298 values between -2.17 and -1.031536 values between -1.03 and 0.111513 values between 0.11 and 1.25158 values between 1.25 and 2.3869 values between 2.38 and 3.5227 values between 3.52 and 4.6615 values between 4.66 and 5.7913 values between 5.79 and 6.9312 values between 6.93 and 8.071 values between 8.07 and 9.212 values between 9.21 and 10.344 values between 10.34 and 11.481 values between 16.03 and 17.171 values between 23.99 and 25.13offnegativepositiveoften

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. long
  2. long
  3. Long
  4. Long
  5. →long
  6. 长
  7. (long
  8. .long
  9. -long
  10. LONG

Suppresses

  1. 骀
  2. ricing
  3. neau
  4. odox
  5. CAA
  6. äter
  7. 鹃
  8. ogn
  9. oge
  10. IONS