Neuron Explorer

Qwen3-0.6B

'fixed' as a finance or cost term

A neuron in layer 4 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 'fixed' in phrases like 'fixed asset', 'fixed amount', 'fixed-income' and 'fixed price', especially in finance and insurance text. Turning it up did not bring out finance-related writing in the sample outputs, so the link is only seen in the examples.

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

    value function for a fixed policy, in a simulated domain.⏎the application examples in programming are very good and

  2. #2 peak

    to interpret this extremely useful financial measurement.⏎Financial ratios such as the fixed asset turnover help financial analysts, management, and investors alike to make critical decisions whether

  3. #3 peak

    calculating the current bond price with our bond valuation calculator, let's take some time to talk about what a bond is. A bond is one of the most prevalent fixed-income securities. When an entity issues bonds, it is considered as acquiring funding from

  4. #4 peak

    company and, usually, to the assets of the company in the event of its dissolution. Dividends are usually set at a fixed annual rate that must be paid before dividends are distributed to the common stockholders.

  5. #5 peak

    to maintain coverage.⏎b. Deductible: The predetermined amount that individuals must pay out-of-pocket before insurance coverage kicks in.⏎c. Copayment: A fixed amount individuals pay for a specific healthcare service, such as a doctor’s visit or

  6. #6 peak

    products and services at similar prices throughout the supply chain.⏎How is this different from the fixed cost? In the case of fixed price, you have to factor in costs of production before you come up with your

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: can I help with this question? I can't help with this question.] Okay, this is a bit of a stretch. I need to think that I can't help with this question.] Okay, I need to tell you to

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

    children learn and develop. There’s no reason why children with autism who use visual supports are any different. So why not bring play and fun games for children with autism into your

  2. #2 peak

    reasonable.⏎But, of course, the reasonable nature of Utterson’s deductions proves precisely their downfall. Utterson remains so adamantly rational and sensible that he never once admits the possibility of a supernatural explanation. He is the embodiment of the Victorian mind

  3. #3 peak

    existence, blazing a trail of death and destruction that prompted mobs of angry villagers to take up torches and pitchforks against him on the silver screen. Never once during his rampage, however, did the monster question the honesty of his ultimate creator

  4. #4 peak

    .⏎This disobedient crowd, momentarily growing as others pushed forward from the main body of people who stayed respectfully at a distance, failed to heed Moses. As soon as Moses reached them, he again shouted his warning.⏎“Move back at once all

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. L2:3033 suppresses
  2. L3:2144 suppresses
  3. L3:1281 weakly suppresses
  4. L1:1401 weakly excites
  5. L2:745 weakly excites
  6. L2:597 weakly excites
  7. L1:515 weakly suppresses 'about' in 'it's not about' phrasing
  8. L3:2736 weakly excites
  9. L1:1287 weakly excites
  10. L3:115 weakly suppresses
  11. L2:2513 weakly suppresses
  12. L2:1708 weakly suppresses
  13. L2:2799 weakly excites
  14. L1:1386 weakly suppresses
  15. L3:456 weakly suppresses
  16. L2:2900 weakly excites
  17. L2:115 weakly excites
  18. L3:1592 weakly excites
  19. L3:464 weakly excites
  20. L0:1345 weakly excites
  21. L0:2320 weakly excites
  22. L0:1277 weakly suppresses
  23. L3:2003 weakly suppresses
  24. L2:2982 weakly suppresses

Feeds later layers

  1. L8:1461 excites
  2. L11:1430 excites
  3. L5:1260 weakly excites
  4. L23:1674 weakly excites
  5. L26:2879 weakly excites
  6. L12:2525 weakly excites
  7. L16:2089 weakly suppresses
  8. L5:946 weakly suppresses
  9. L15:842 weakly suppresses 'not' in 'if not most/all' phrases
  10. L19:2846 weakly suppresses
  11. L8:1167 weakly excites
  12. L5:1554 weakly suppresses
  13. L11:2468 weakly excites
  14. L21:225 weakly excites
  15. L5:1947 weakly suppresses
  16. L12:54 weakly excites
  17. L15:1114 weakly excites
  18. L16:921 weakly suppresses
  19. L16:2848 weakly excites
  20. L20:953 weakly excites
  21. L17:2463 weakly excites
  22. L25:694 weakly suppresses
  23. L11:812 weakly suppresses
  24. L6:1024 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 -1.78 and -1.701 values between -1.70 and -1.611 values between -0.91 and -0.821 values between -0.73 and -0.653 values between -0.65 and -0.563 values between -0.56 and -0.478 values between -0.47 and -0.3822 values between -0.38 and -0.3086 values between -0.30 and -0.21315 values between -0.21 and -0.12872 values between -0.12 and -0.031416 values between -0.03 and 0.05755 values between 0.05 and 0.14315 values between 0.14 and 0.2378 values between 0.23 and 0.3219 values between 0.32 and 0.403 values between 0.40 and 0.492 values between 0.49 and 0.581 values between 0.58 and 0.671 values between 0.75 and 0.841 values between 0.93 and 1.021 values between 1.10 and 1.191 values between 1.28 and 1.371 values between 1.45 and 1.541 values between 1.63 and 1.72offnegativepositiveoften

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. 工委
  2. 탕
  3. Scanner
  4. aneously
  5. раг
  6. _fixed
  7. .fixed
  8. licensee
  9. -fixed
  10. ArrayOf

Suppresses

  1. FG
  2. velt
  3. зем
  4. ifique
  5. ém
  6. SV
  7. .geom
  8. 括
  9. �
  10. thế