Neuron Explorer

Qwen3-0.6B

"that" as a pointing word

A neuron in layer 6 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 "that" when it points back to something just mentioned, as in "that day" or "that work." Turning it up did not make the model's text show this pattern, so the steering is unclear.

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

    also be rocky and have a composition much like Earth’s. Some might even harbor life. Over 99% of those alien worlds stay undiscovered — however that is about to vary.⏎With NASA’s new exoplanet-hunter house telescope

  2. #2 peak

    time throughout the summer to learn and complete work; even just 15-30 minutes a day will allow students to keep their skills sharp and knock out that summer work so they can enjoy their break worry-free.⏎How can you take this

  3. #3 peak

    agency.⏎Although they were in a hurry to complete the structure, the builders halted the work on Sunday for a time of worship and Bible study. It happened on that day that a Ho-Chunk man named Four Legs (a different individual than the

  4. #4 peak

    a very important gap for students experiencing race-related stress<|endoftext|>Western tradition gets upheld as a lone bastion of freedom, the shining city on a hill. To make that

  5. #5 peak

    under even greater strain. To help ease that pressure, Bruce and other scientists are working to understand and help improve control of crop

  6. #6 peak

    and is in part driven by the ‘diabetes epidemic’.⏎... the ‘healthier world of tomorrow’ requires that scientists today build the knowledge base to create that ’healthier world of tomorrow’.⏎Equal access to a healthy lifestyle⏎A

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 question about the problem, and I have a problem with the problem. The problem is not only about the problems, and the problems are in the same with each other. The problem is not just about the problems. I

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

    is an element , where is some object of , which exhibits representability of via the Yoneda lemma. That is, any element induces, in natural bijective fashion, a natural transformation⏎and is universal if is an isomorphism.⏎

  2. #2 peak

    . If the ingredient list is not available on the product, check<|endoftext|>Help support New Advent and get the full contents of this website as an instant download or CD-ROM. Includes the Catholic Encyclopedia, Church Fathers, Summa, Bible and more all

  3. #3 peak

    of Arts in Kyoto, Japan. This provides evidence that exposure to certain patterns of speech can influence one's perceptions of musical rhythms. In

  4. #4 peak

    little to salvage for homeowners caught in harm's way.⏎Climate change will disproportionately affect California’s most disadvantaged and isolated communities, according to a recent report from the Pacific Institute.⏎By looking at a broad array of factors – from social indicators such as income

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. L3:1003 excites
  2. L4:851 excites
  3. L4:468 excites "that" after "than" or "like"
  4. L4:2051 weakly suppresses
  5. L0:2841 weakly suppresses
  6. L5:2747 weakly suppresses
  7. L1:2213 weakly suppresses
  8. L1:875 weakly excites
  9. L4:1762 weakly suppresses "this" in time and context phrases
  10. L2:2259 weakly excites
  11. L5:1537 weakly excites
  12. L5:1561 weakly suppresses
  13. L4:1042 weakly suppresses
  14. L4:1748 weakly excites
  15. L5:1446 weakly suppresses
  16. L2:2794 weakly suppresses
  17. L2:2850 weakly excites
  18. L2:2615 weakly suppresses
  19. L1:2172 weakly suppresses
  20. L1:1036 weakly suppresses
  21. L1:146 weakly excites
  22. L3:2518 weakly suppresses
  23. L3:81 weakly excites
  24. L0:2204 weakly excites

Feeds later layers

  1. L10:965 excites the word 'that' introducing a clause
  2. L15:470 excites the word 'this' after a preposition
  3. L12:521 excites
  4. L7:1600 weakly excites
  5. L16:715 weakly excites
  6. L18:814 weakly suppresses
  7. L24:668 weakly excites
  8. L11:2338 weakly suppresses
  9. L17:655 weakly excites
  10. L18:1393 weakly suppresses
  11. L14:2489 weakly excites
  12. L17:221 weakly excites
  13. L8:1182 weakly suppresses
  14. L16:896 weakly excites
  15. L19:2547 weakly excites
  16. L13:1007 weakly excites
  17. L16:2861 weakly excites
  18. L21:323 weakly excites
  19. L25:274 weakly excites
  20. L13:2397 weakly suppresses
  21. L13:1376 weakly suppresses
  22. L25:2529 weakly excites
  23. L10:2276 weakly excites
  24. L11:810 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.

1 values between -1.26 and -1.012 values between -1.01 and -0.765 values between -0.76 and -0.5197 values between -0.51 and -0.261739 values between -0.26 and -0.011950 values between -0.01 and 0.2382 values between 0.23 and 0.4814 values between 0.48 and 0.734 values between 0.73 and 0.984 values between 1.23 and 1.481 values between 1.97 and 2.221 values between 2.22 and 2.471 values between 3.22 and 3.461 values between 4.46 and 4.711 values between 4.95 and 5.202 values between 5.45 and 5.701 values between 6.44 and 6.691 values between 6.94 and 7.191 values between 8.43 and 8.68offnegativepositiveoften

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. sett
  2. 坊
  3. omm
  4. opr
  5. hc
  6. raj
  7. hh
  8. der
  9. 结论
  10. hed

Suppresses

  1. -mile
  2. -opacity
  3. _each
  4. -fontawesome
  5. -average
  6. ctica
  7. cente
  8. -density
  9. �权
  10. each