Neuron Explorer

Qwen3-0.6B

'from' after 'Retrieved' and source phrases

A neuron in layer 5 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 'from' in citation phrases like 'Retrieved from', and in phrases like 'can be purchased from' or 'expected too much from'. It goes negative on 'by' after passive verbs such as 'acquired by' or 'bought by'. Turning it up or down did not produce a clear change in the writing.

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 Publication⏎Blanks, C. M. (1948). San Antonio the Inland Gate. Retrieved from https://digitalcommons.pvamu.edu/pvamu-theses/39

  2. #2 peak

    eating disorders. The trait of perfectionism is common among people with obsessive-compulsive disorders. Or it could be a response to having parents who expected too much from you. Maybe they never let you off the hook, even if you got

  3. #3 peak

    6C>.⏎- APA: Learning Math; Measurement: Indirect Measurement And Trigonometry. Boston, MA: WGBH Media Library & Archives. Retrieved from http://openvault.wgbh.org/catalog/V_4E4EE3

  4. #4 peak

    although some states also award goods or services. In many cases, a winner must pay taxes on the prize amount before receiving it.⏎Lottery tickets can be purchased from retail stores, newsstands, gas stations, and convenience stores. Some lotter

  5. #5 peak

    1853-1863). International Journal on Stereo & Immersive Media, 2(2), 54-69. Retrieved from https://revistas.ulusofona.pt/index.php/stereo/article/view/

  6. #6 peak

    ]. (2018, July 6). 1929 Stock Market Crash and the Great Depression –⏎Documentary [Video File]. Retrieved from https://youtu.be/qlSxPouPCIM⏎The post

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: there is a 200-gram mass of 10 grams. The mass of the object is directly proportional to the product of the kinetic energy, which is 200 grams. The product of the kinetic energy is

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 closed to the public. It is undergoing a state-ordered cleanup for soil and groundwater contamination linked to Peerless Photo Products, a previous owner that was acquired by Agfa.⏎A security guard who never heard of Nikola Tesla turned away visitors on

  2. #2 peak

    sold to a private owner.⏎Over the years, the hotel changed hands several times and underwent several renovations. In 2007, the hotel was purchased by the French Lick Resort and underwent a $500 million restoration, restoring

  3. #3 peak

    is not a teacher. Coleman worked for McGraw-Hill (textbook publishing company) before leaving to create GROW a curriculum testing company which was later bought by McGraw-Hill. Bill Gates (who funded the creation of Common Core)

  4. #4 peak

    , a tag library defines a set of Java classes for the custom JSP actions that the tag library supports. Invoking the tag library imports custom processing and actions into the Java environment so a Java runtime environment like JRun 3.0 will

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. L3:822 suppresses
  2. L4:1038 weakly excites
  3. L0:2358 weakly suppresses
  4. L0:1923 weakly suppresses
  5. L4:97 weakly suppresses
  6. L3:3055 weakly excites
  7. L2:2661 weakly suppresses
  8. L2:2895 weakly suppresses
  9. L4:2050 weakly excites
  10. L2:1352 weakly suppresses
  11. L3:1568 weakly suppresses
  12. L4:557 weakly suppresses
  13. L3:2005 weakly excites
  14. L1:2846 weakly excites
  15. L3:1694 weakly suppresses
  16. L4:1662 weakly suppresses
  17. L4:1528 weakly suppresses
  18. L4:2325 weakly suppresses
  19. L0:1797 weakly excites
  20. L3:1342 weakly excites
  21. L0:1839 weakly excites
  22. L4:12 weakly suppresses
  23. L1:1720 weakly suppresses
  24. L3:2375 weakly suppresses

Feeds later layers

  1. L9:2267 excites 'with' after exchange nouns
  2. L21:2262 weakly excites
  3. L9:654 weakly excites
  4. L17:2984 weakly excites blank before a number in prevalence claims
  5. L22:2344 weakly excites number after 'account for' or 'over'
  6. L15:2416 weakly suppresses
  7. L7:1393 weakly excites
  8. L21:1203 weakly excites
  9. L6:2695 weakly suppresses
  10. L19:2244 weakly excites
  11. L17:978 weakly excites
  12. L7:1982 weakly suppresses
  13. L25:1488 weakly excites
  14. L14:725 weakly excites
  15. L8:1034 weakly excites
  16. L27:2847 weakly excites
  17. L9:2461 weakly excites
  18. L6:1834 weakly excites
  19. L17:2711 weakly excites
  20. L18:3001 weakly excites
  21. L9:186 weakly suppresses
  22. L7:1028 weakly excites
  23. L15:2463 weakly suppresses
  24. L7:1326 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 -1.86 and -1.761 values between -1.65 and -1.542 values between -1.00 and -0.894 values between -0.89 and -0.783 values between -0.78 and -0.681 values between -0.68 and -0.5713 values between -0.57 and -0.4629 values between -0.46 and -0.3578 values between -0.35 and -0.25293 values between -0.25 and -0.141027 values between -0.14 and -0.031649 values between -0.03 and 0.08569 values between 0.08 and 0.19169 values between 0.19 and 0.3037 values between 0.30 and 0.408 values between 0.40 and 0.516 values between 0.51 and 0.621 values between 0.62 and 0.732 values between 0.73 and 0.842 values between 0.84 and 0.941 values between 0.94 and 1.051 values between 1.16 and 1.274 values between 1.27 and 1.381 values between 1.38 and 1.481 values between 1.48 and 1.591 values between 1.59 and 1.701 values between 1.70 and 1.811 values between 1.92 and 2.021 values between 2.13 and 2.241 values between 2.35 and 2.46offnegativepositiveoften

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. mitted
  2. 用户提供
  3. .Generation
  4. 屏幕上
  5. 更多信息
  6. 媒体报道
  7. 官方微信
  8. 层出
  9. nią
  10. 覆盖率

Suppresses

  1. uld
  2. 降落
  3. oard
  4. dress
  5. 竿
  6. adh
  7. dsl
  8. scoped
  9. ë
  10. rious