Neuron Explorer

Qwen3-0.6B

'a' before an object noun

A neuron in layer 20 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

Fires on the word 'a' when it comes right after a verb or preposition and a noun is about to follow, such as 'when requesting a correction'. Turning it down made the model repeat 'a a a', so the output is mostly broken.

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

    VECE20|⏎|Order Information:||Web: http://www.tandfonline.com/pricing/journal/VECE20|⏎When requesting a correction, please mention this item's handle: RePEc:taf:j

  2. #2 peak

    objective.⏎Student data on learning objectives also generate a learning estimate that appears in the instructor's Learning Dashboard. The learning estimate predicts how well a student will perform on a learning objective.⏎Summative assessments are assignments that assess mastery of material. Scores

  3. #3 peak

    adjustments to the apparatus are being made, the hood should be kept closed, with vertical sashes down and horizontal sashes closed, to help prevent the spread of a fire, spill, or other hazard into the laboratory. Sliding sashes should

  4. #4 peak

    based on the borrower's income, expenses, and credit history. The results of a stress test can impact the amount a borrower is able to borrow for a mortgage and the interest rate they receive.⏎Debt Servicing Ratio is a measure

  5. #5 peak

    in this study are freely available to the broader research community, and researchers can use these data to study the cells in greater depth and determine how phosphorylation events determine a cell's fate," said Dr. Laurence M. Brill, senior scientist at

  6. #6 peak

    Angle in End Milling⏎End mills are used to form shapes and holes during milling and can cut a variety of materials in several directions. The end mill is generally described as a positive

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: a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a a

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

    prove its identity as a unique physiographic region and as a region that is beneficial to grape growing. Some of the criteria used by the federal government to determine whether an area meets this standard include distinct climate, soil type, and elevation and other

  2. #2 peak

    incident to two nodes. We define adjacent nodes of a given node as alternate nodes of incident edges. Adjacent nodes are sometimes referred to as neighbors. The degree of a node is the number of its incident edges or

  3. #3 peak

    the exact value of the number. Once you have the answer, you can move onto the next step in interpreting the result.⏎The first step in finding the percentage of a

  4. #4 peak

    percent. It’s easy to do this with the calculator. You can use a different method to find the percentage of a number.⏎Another way to calculate the percentage of a number is to divide the number by its maximum possible value. You can do

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. L19:2972 suppresses
  2. L19:684 suppresses
  3. L18:692 excites
  4. L19:2269 suppresses
  5. L19:654 excites
  6. L18:1538 suppresses
  7. L18:3050 suppresses
  8. L17:1527 excites
  9. L18:2323 suppresses
  10. L18:2947 excites
  11. L18:2404 suppresses
  12. L17:817 suppresses
  13. L16:2332 suppresses
  14. L19:1352 suppresses
  15. L14:2629 suppresses
  16. L19:2274 excites
  17. L18:376 weakly suppresses
  18. L16:2363 weakly excites
  19. L19:626 weakly excites
  20. L19:1710 weakly excites
  21. L18:406 weakly excites
  22. L17:1429 weakly excites
  23. L19:1305 weakly suppresses unclear: verbs after a noun or 'that'
  24. L0:1826 weakly excites

Feeds later layers

  1. L22:103 strongly excites
  2. L25:45 strongly excites
  3. L27:797 strongly suppresses
  4. L21:1457 strongly excites
  5. L25:1586 strongly suppresses
  6. L27:663 strongly suppresses
  7. L24:823 strongly suppresses
  8. L26:484 strongly suppresses
  9. L24:2538 strongly suppresses
  10. L27:709 suppresses
  11. L25:1132 excites
  12. L26:213 excites
  13. L27:231 excites
  14. L24:488 excites
  15. L25:871 suppresses
  16. L27:1044 suppresses
  17. L21:1720 excites
  18. L27:173 excites
  19. L21:775 excites the article 'a' before a noun
  20. L27:2396 suppresses
  21. L23:1220 weakly suppresses
  22. L23:119 weakly suppresses
  23. L21:592 weakly excites
  24. L25:568 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 -14.69 and -14.072 values between -14.07 and -13.461 values between -11.61 and -10.993 values between -10.99 and -10.371 values between -9.76 and -9.141 values between -9.14 and -8.531 values between -8.53 and -7.911 values between -7.91 and -7.291 values between -7.29 and -6.681 values between -6.68 and -6.064 values between -6.06 and -5.441 values between -5.44 and -4.835 values between -4.83 and -4.212 values between -4.21 and -3.6011 values between -3.60 and -2.988 values between -2.98 and -2.3620 values between -2.36 and -1.7540 values between -1.75 and -1.13123 values between -1.13 and -0.522728 values between -0.52 and 0.10812 values between 0.10 and 0.7265 values between 0.72 and 1.3325 values between 1.33 and 1.9516 values between 1.95 and 2.5712 values between 2.57 and 3.186 values between 3.18 and 3.803 values between 3.80 and 4.422 values between 4.42 and 5.032 values between 5.03 and 5.654 values between 5.65 and 6.263 values between 6.26 and 6.882 values between 8.11 and 8.731 values between 9.35 and 9.96offnegativepositiveoften

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. esse
  2. inished
  3. icated
  4. 各
  5. imer
  6. each
  7. -Saharan
  8. rete
  9. IVED
  10. ihan

Suppresses

  1. a
  2. an
  3. 一个
  4. 一种
  5. →a
  6. something
  7. 一项
  8. 一款
  9. 一種
  10. 的一种