Neuron Explorer

Qwen3-0.6B

'as well as' and list connectors

A neuron in layer 18 of Qwen3-0.6B. Its strongest snippets are mixed: they peak on different tokens. It is featured, so it has steering outputs.

What this neuron does

This neuron fires on the word 'as' after 'as well', and on 'and' in lists of things being compared or added together. It is a weak fit overall. Turning it up produced repeated phrases like 'a box of any other product'.

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

    similar to Depart.⏎English → Indonesian⏎English → English⏎v 1: move away from a place into another direction; “

  2. #2 peak

    is still practiced within the Vietnamese and French adapted systems, as well as through the teachings of Jeffrey Yuen. The channel systems are considered by many to

  3. #3 peak

    in our nation, health care system and medical research, which does not have enough representation from the Black community.”⏎The report details the disparities for African Americans as well as many other groups, highlights areas where some progress has been made, and lays out

  4. #4 peak

    012 hurricane season, the industry continues to build upon critical lessons learned from 2008's major hurricanes, Gustav and Ike, as well as other powerful storms, such as 2005's Katrina and Rita and

  5. #5 peak

    reading instructions during the test, right? Also, this acquired familiarity is going to help you keep calm during the test. For this writing and language section as well as the others.⏎So, without further ado, let’s go over the three broad

  6. #6 peak

    Zimmer (1985), and Kaufman (1990), and some highlights are in Roberson (1980, 1986, 1993). The Kaufman book — Field Guide to

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: the number of people who have been in the number of people in the number of people in the number of people in the number of people in the number of people in the number of people in the number of people in the number of people in

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

    and useable - the rest is just science.⏎Since its 2003 launch, Drift Catchers have been used in 27 projects in nine states by scores of trained volunteers and community leaders. Below are two case studies.⏎

  2. #2 peak

    -income mothers next week. They will be invited to join the study, which is called Baby’s First Years, shortly after giving birth at one of ten hospitals in four cities across the United States (to avoid influencing the experiment, the researchers asked The

  3. #3 peak

    service is advertised as providing a large computing capacity (potentially many servers) much faster and cheaper than building a physical server farm.⏎AWS is located in 10 geographical "regions": US East (Northern Virginia), where the majority of AWS servers

  4. #4 peak

    queries, or even monitor employees. To understand people’s trust and attitudes towards workplace AI, we surveyed over 17,000 people from 17 countries: Australia, Brazil, Canada, China, Estonia, Finland, France,

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. L16:1430 suppresses
  2. L17:2052 suppresses
  3. L15:2067 excites
  4. L16:1782 suppresses
  5. L16:2962 suppresses unclear: weak mid-phrase prepositions
  6. L16:1361 suppresses
  7. L15:1738 excites
  8. L14:2756 suppresses
  9. L17:1470 excites
  10. L15:292 excites
  11. L16:1733 excites
  12. L13:2083 suppresses
  13. L12:2832 suppresses the word 'even' as emphasis
  14. L14:2360 excites
  15. L17:1009 excites
  16. L17:586 suppresses
  17. L14:980 suppresses
  18. L12:115 suppresses
  19. L15:1693 suppresses
  20. L17:1017 suppresses
  21. L17:1079 suppresses
  22. L16:1240 suppresses
  23. L17:1151 excites
  24. L17:2264 weakly excites

Feeds later layers

  1. L19:2484 suppresses
  2. L19:354 suppresses
  3. L19:666 suppresses
  4. L19:1640 excites
  5. L19:2756 weakly excites
  6. L20:1923 weakly suppresses
  7. L19:356 weakly suppresses
  8. L19:519 weakly suppresses
  9. L19:331 weakly suppresses
  10. L19:2897 weakly suppresses
  11. L22:870 weakly excites
  12. L19:2298 weakly suppresses
  13. L19:355 weakly suppresses
  14. L20:2629 weakly suppresses
  15. L20:1189 weakly excites
  16. L22:2904 weakly excites
  17. L19:2204 weakly excites
  18. L19:1861 weakly suppresses
  19. L19:2274 weakly excites
  20. L19:416 weakly suppresses
  21. L21:1931 weakly suppresses
  22. L20:708 weakly suppresses
  23. L21:1126 weakly excites
  24. L21:2287 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 -5.77 and -5.411 values between -5.41 and -5.065 values between -4.70 and -4.356 values between -4.35 and -3.998 values between -3.64 and -3.289 values between -3.28 and -2.9312 values between -2.93 and -2.5732 values between -2.57 and -2.2227 values between -2.22 and -1.8662 values between -1.86 and -1.51110 values between -1.51 and -1.15202 values between -1.15 and -0.80364 values between -0.80 and -0.44862 values between -0.44 and -0.081706 values between -0.08 and 0.27232 values between 0.27 and 0.6398 values between 0.63 and 0.9856 values between 0.98 and 1.3444 values between 1.34 and 1.6922 values between 1.69 and 2.059 values between 2.05 and 2.4015 values between 2.40 and 2.764 values between 2.76 and 3.113 values between 3.11 and 3.475 values between 3.47 and 3.823 values between 3.82 and 4.183 values between 4.18 and 4.532 values between 4.53 and 4.891 values between 4.89 and 5.241 values between 5.24 and 5.601 values between 5.60 and 5.951 values between 6.31 and 6.661 values between 8.08 and 8.44offnegativepositiveoften

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. bserv
  2. TestMethod
  3. grey
  4. ason
  5. ọi
  6. edian
  7. 彩
  8. nev
  9. 任何
  10. 个工作

Suppresses

  1. ature
  2. 本身
  3. مباشر
  4. מונה
  5. отов
  6. HL
  7. ologically
  8. ogn
  9. NAT
  10. 牯