Neuron Explorer

Qwen3-0.6B

the last digit of decade years

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 lights up on the final zero in decade years like 1970s, 1980s or 1930s, so it is reacting to the point in a sentence where a decade is named. Turning it up pushed the model toward unrelated maths word problems and did not make it talk about history.

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

    and Quantum Communication.⏎How it started:⏎Artificial Intelligence started in the 1950s, but its importance was recognized in the 1970s.⏎Japan was the first to take this initiative and in 198

  2. #2 peak

    mass incarceration of Japanese Americans, Asian Americans called for affirmative action programs for racially excluded minorities. From the 1960s through the 1980s, Asian Americans helped force higher education open for excluded communities of color and benefited

  3. #3 peak

    insights into the Martian past demonstrated the science potential of investing in imaging of the planets from orbit. By the end of their missions in the early 1980s, the Viking landers and orbiters had accumulated over 50,

  4. #4 peak

    resulted in a period when nearly every American family had a shot at a better life.⏎Our recent history tells a different story. Since the early 1980

  5. #5 peak

    ⏎It became particularly strong during World War II--when advancing Japanese forces cut off U.S. access to hevea--and during the 1970s.⏎Guayule grows much faster than hevea, but the latter

  6. #6 peak

    Norman Davies noted that American forces fought with distinction on two fronts, but that it was our home factories that turned the tide. In the late 1930's, American industry produced consumer goods at a dog's pace. By 1

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 lot of things." The number of the blog post is a great place for the word 'something' in the context of a blog post. The first thing to note is that the word 'something' is used in a wide range

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

    etal force acting on a car of mass 2000 Kg turning around a circular bend of radius 50 meters at the constant speed of 36 km/hr?⏎A) 4,000 N⏎B)

  2. #2 peak

    ) with a reasonable profit at the present price of crude oil, and 2,500 barrels per acre for an area of 2,248 acres (5,620,000 barrels total), with no

  3. #3 peak

    optical mice as well as 800 DPI laser mice are very popular, but some high-end models are capable of 1000, 1600 or even 2000 DPI tracking speeds.⏎The Personal System

  4. #4 peak

    drop for different weight bullets ranging from 250, 325, and 405 grains.⏎This following table is a ballistics

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. L5:1641 suppresses
  2. L5:1072 weakly excites
  3. L4:12 weakly suppresses
  4. L4:2492 weakly excites
  5. L4:2888 weakly suppresses
  6. L5:881 weakly excites
  7. L4:245 weakly excites
  8. L2:1605 weakly suppresses
  9. L2:208 weakly excites
  10. L3:612 weakly excites
  11. L1:1254 weakly excites
  12. L5:2154 weakly excites
  13. L5:28 weakly excites
  14. L4:1996 weakly excites
  15. L1:462 weakly suppresses
  16. L3:2980 weakly suppresses
  17. L0:2023 weakly suppresses
  18. L3:1678 weakly suppresses
  19. L5:2513 weakly excites
  20. L5:2937 weakly excites
  21. L5:2727 weakly excites
  22. L2:2330 weakly suppresses
  23. L3:512 weakly suppresses
  24. L3:176 weakly suppresses

Feeds later layers

  1. L7:1808 strongly excites
  2. L24:1676 excites zeros in large round numbers
  3. L27:2476 weakly suppresses
  4. L17:409 weakly excites
  5. L8:1402 weakly excites
  6. L13:2379 weakly excites the digit zero in numbers
  7. L9:977 weakly excites the last digit of a decade year
  8. L22:687 weakly suppresses
  9. L20:854 weakly suppresses
  10. L17:2270 weakly suppresses
  11. L15:1965 weakly suppresses
  12. L14:2970 weakly excites
  13. L25:2616 weakly suppresses
  14. L19:623 weakly suppresses
  15. L7:1696 weakly excites
  16. L21:474 weakly suppresses
  17. L22:2300 weakly suppresses
  18. L27:2740 weakly suppresses
  19. L17:1466 weakly excites
  20. L26:847 weakly suppresses
  21. L13:2708 weakly suppresses
  22. L17:2305 weakly suppresses
  23. L10:1792 weakly suppresses
  24. L23:2406 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.

2 values between -2.31 and -2.151 values between -1.48 and -1.311 values between -1.31 and -1.142 values between -1.14 and -0.981 values between -0.98 and -0.813 values between -0.81 and -0.644 values between -0.64 and -0.4747 values between -0.47 and -0.31434 values between -0.31 and -0.142072 values between -0.14 and 0.031178 values between 0.03 and 0.20130 values between 0.20 and 0.367 values between 0.36 and 0.533 values between 0.53 and 0.704 values between 0.70 and 0.863 values between 0.86 and 1.032 values between 1.03 and 1.202 values between 1.20 and 1.371 values between 1.37 and 1.532 values between 1.53 and 1.701 values between 1.70 and 1.872 values between 1.87 and 2.031 values between 2.20 and 2.371 values between 2.37 and 2.541 values between 2.87 and 3.041 values between 3.71 and 3.872 values between 4.21 and 4.38offnegativepositiveoften

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. awning
  2. awns
  3. ��
  4. quartered
  5. ��
  6. licted
  7. elps
  8. assador
  9. チェ
  10. ��

Suppresses

  1. да
  2. атель
  3. ale
  4. опы
  5. Lock
  6. immer
  7. дал
  8. 使用
  9. Use
  10. 输卵管