Neuron Explorer

Qwen3-0.6B

"from" after a change in amount

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 "from" when it names the starting value of a change, as in "doubled from 3.5 to" or "increased by 62%, from 47.3 years." Turning it up did not produce a clear theme in the replies.

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

    1980 and 2010, the average years of schooling for females aged 15–19 in MENA more than doubled from 3.5 to

  2. #2 peak

    damaging groves. But this year the industry is "making a comeback", according to Ocean Spray. The group forecasts the Florida grapefruit harvest will almost double from 19 million boxes a year ago to 28 million this year. The

  3. #3 peak

    rise over the past three and a half decades. The average age of first-time mothers has increased by 3.6 years over last 36 years, from the age of 21.4 years in 1970 to

  4. #4 peak

    grow from 6Gbps to 10Gbps. The net result of these simultaneous forces is a 25% increase in monthly transit fees, from $16,000 per month to $20,00

  5. #5 peak

    alcohol and marijuana use changed little from 2009 to 2012. However, lifetime RR for use of other illicit drugs increased substantially, from an RR of 1.8 (95% CI, 1.

  6. #6 peak

    artners are using the iPad to learn and study.⏎And the AP found that the percentage of preschoolers with a social media account increased by almost a third, from about one in ten to about one-half of one percent.⏎Here’s what you

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 3-digit numbers with a second digit of 3 and first digits of 1 and 2 and 3, and 4, 5, 6, 7, 8, 9, 1

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

    to health. It wasn’t long, however, before new drugs were developed that rendered this treatment obsolete.⏎As such, the building’s role would shift several times over the next six decades before it was added to the National Register of Historic Places in

  2. #2 peak

    that the fractional red-shift d / is constant throughout a given spectrum. Each wave-length is increased, and, consequently, the energy in each quantum is reduced, by the constant factor 1 + d / . We might naturally conclude that the total

  3. #3 peak

    student blogs were nothing more than a class journal posted online. They were digital notebooks. However, over time, they became a place for student voice. They evolved as my thinking about blogging evolved. So, here is what I’ve learned along the

  4. #4 peak

    compromise responses to climate changes – both ‘normal’ variation and shocks as well as new variability and increased uncertainty.⏎Third, local support systems are changing, and so is people’s ability to make use of them. The report shows that changes in locally

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. L3:754 suppresses
  2. L2:1832 weakly suppresses
  3. L3:1942 weakly suppresses
  4. L4:2080 weakly excites
  5. L4:2337 weakly excites
  6. L5:46 weakly excites
  7. L5:27 weakly excites
  8. L4:1356 weakly excites
  9. L5:57 weakly suppresses
  10. L4:1308 weakly suppresses
  11. L5:2530 weakly excites
  12. L1:1971 weakly suppresses
  13. L4:1992 weakly excites
  14. L4:1262 weakly suppresses
  15. L1:2801 weakly excites
  16. L1:2650 weakly suppresses
  17. L4:1238 weakly suppresses
  18. L2:78 weakly suppresses
  19. L4:2877 weakly suppresses
  20. L3:713 weakly excites
  21. L5:2216 weakly excites
  22. L5:2398 weakly excites
  23. L0:2760 weakly suppresses
  24. L3:342 weakly excites

Feeds later layers

  1. L11:2662 excites
  2. L12:1467 excites numbers after dates and figures
  3. L7:2592 excites
  4. L9:743 excites
  5. L8:2924 excites
  6. L17:2732 excites
  7. L18:1660 excites
  8. L8:229 excites
  9. L8:2018 excites
  10. L14:2475 excites
  11. L11:2970 weakly excites
  12. L13:2224 weakly excites
  13. L14:867 weakly excites
  14. L7:582 weakly excites
  15. L9:1257 weakly excites
  16. L12:1294 weakly excites
  17. L13:2502 weakly excites
  18. L7:2210 weakly excites
  19. L20:2968 weakly suppresses 'more' or 'not' before a likelihood claim
  20. L13:2958 weakly excites
  21. L11:1958 weakly suppresses
  22. L23:1518 weakly suppresses
  23. L7:213 weakly excites
  24. L8:1401 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.

2 values between -0.97 and -0.844 values between -0.84 and -0.701 values between -0.70 and -0.5720 values between -0.57 and -0.4356 values between -0.43 and -0.30307 values between -0.30 and -0.161334 values between -0.16 and -0.031683 values between -0.03 and 0.11407 values between 0.11 and 0.2471 values between 0.24 and 0.3813 values between 0.38 and 0.513 values between 0.51 and 0.652 values between 0.65 and 0.782 values between 0.78 and 0.921 values between 1.05 and 1.191 values between 2.00 and 2.131 values between 4.30 and 4.43offnegativepositiveoften

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. 换成
  2. @update
  3. 改动
  4. 改为
  5. 改成
  6. 变为
  7. 替换
  8. 修订
  9. đổi
  10. 更改

Suppresses

  1. ết
  2. щен
  3. 넷
  4. _rng
  5. رب
  6. fähig
  7. nox
  8. uncan
  9. 甥
  10. unes