Neuron Explorer

Qwen3-0.6B

'he' in pheasant, 'mouth' in Dartmouth

A neuron in layer 1 of Qwen3-0.6B. Its strongest snippets are fairly consistent: several of them peak on the same token. It is featured, so it has steering outputs.

What this neuron does

This neuron responds to parts of words that are split up, like the 'he' in 'pheasant' and 'mouth' in 'Dartmouth'. Turning it up or down did not give an obvious, consistent change.

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

    include: Oliphant, Olifant, Olifard and others.⏎First found in Perthshire (Gaelic: Siorrachd Pheairt) former county in the present day Council Area of Perth and Kinross

  2. #2 peak

    - Great Hornbill (Buceros bicornis),⏎- Rufous-necked Hornbill (Aceros nipalensis),⏎- Grey Peacock Pheasant (Polyplectron bicalcaratum),⏎

  3. #3 peak

    pheasant) is literally 'mountain bird', and ashibiki is a pillow-word for mountain, which is itself the first half of the word for pheasant.<|endoftext|>North American freshwater fishes are going extinct at an alarming rate compared with

  4. #4 peak

    Pittsburgh Research Laboratory, Pittsburgh, PA 15236⏎Other Occupational Concerns⏎Longwall USA, International Exhibition & Conference<|endoftext|>Hypertext Markup Language. The standard language used to create Web pages. Originally based on SGML,

  5. #5 peak

    – could steer sexual behavior development in offspring. It’s even conceivable that taking antihistamines or pain relievers during pregnancy could play a role.⏎“We’re really interested in the fundamental mechanisms that drive brain development and sex-specific brain development, and

  6. #6 peak

    Old Japan (The Hyakunin-isshu), tr. by William N. Porter, , at sacred-texts.com⏎LONG is the mountain pheasant's tail⏎That curves down in its flight;⏎But longer still, it

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: I have been working on this project, and I have been thinking about it. But I have been working on it for a while, and I have been thinking about it for a while. It has been a long time. It has been

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

    registered user you can vote on this brain teaser, keep track of⏎which ones you have seen, and even make your own.⏎Back to Top<|endoftext|>Dartmouth researchers have developed a protocol that permits cells harvested from melanoma tumors in mice to

  2. #2 peak

    iers’ Protect Greenland’s Landscapes⏎A story by LiveScience published by NBC News features a new study by Lee Corbett, a doctoral student in Dartmouth’s Department of Earth Sciences, that reveals how “ghost glaciers” have protected Greenland

  3. #3 peak

    ⏎A University of Massachusetts Dartmouth education is a foundation from which graduates continually engage and impact their community, both locally

  4. #4 peak

    a model system," said Constance E. Brinckerhoff, PhD, professor of Medicine and of Biochemistry at the Geisel School of Medicine at Dartmouth and a member of the Norris Cotton Cancer Center (NCCC) Mechanism Research

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. L0:1301 suppresses
  2. L0:1224 suppresses
  3. L0:178 suppresses
  4. L0:2308 weakly excites
  5. L0:1 weakly suppresses
  6. L0:581 weakly excites
  7. L0:799 weakly suppresses
  8. L0:2316 weakly suppresses
  9. L0:2621 weakly excites
  10. L0:1999 weakly excites
  11. L0:489 weakly excites
  12. L0:1000 weakly excites
  13. L0:654 weakly suppresses
  14. L0:15 weakly excites
  15. L0:33 weakly excites
  16. L0:3005 weakly excites
  17. L0:1453 weakly suppresses
  18. L0:1102 weakly excites
  19. L0:1837 weakly excites
  20. L0:40 weakly excites 'of' in 'sort of' hedges
  21. L0:2047 weakly excites
  22. L0:1634 weakly suppresses
  23. L0:2321 weakly suppresses
  24. L0:1761 weakly suppresses

Feeds later layers

  1. L3:1955 weakly excites
  2. L19:1312 weakly suppresses
  3. L14:3022 weakly suppresses
  4. L16:1740 weakly suppresses
  5. L21:992 weakly suppresses
  6. L14:2361 weakly excites
  7. L22:2083 weakly suppresses
  8. L7:2185 weakly suppresses
  9. L26:2465 weakly suppresses
  10. L6:2923 weakly excites
  11. L9:1355 weakly excites
  12. L15:167 weakly excites
  13. L26:651 weakly suppresses
  14. L16:1324 weakly suppresses
  15. L21:2113 weakly suppresses
  16. L18:2277 weakly suppresses
  17. L22:1129 weakly excites
  18. L12:882 weakly excites
  19. L6:623 weakly suppresses
  20. L20:2509 weakly suppresses
  21. L20:1746 weakly suppresses
  22. L18:229 weakly excites
  23. L14:2592 weakly excites
  24. L22:2390 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 -0.50 and -0.481 values between -0.48 and -0.463 values between -0.46 and -0.441 values between -0.44 and -0.421 values between -0.42 and -0.394 values between -0.39 and -0.376 values between -0.37 and -0.352 values between -0.35 and -0.336 values between -0.33 and -0.305 values between -0.30 and -0.284 values between -0.28 and -0.2614 values between -0.26 and -0.2423 values between -0.24 and -0.2118 values between -0.21 and -0.1927 values between -0.19 and -0.1733 values between -0.17 and -0.1559 values between -0.15 and -0.12102 values between -0.12 and -0.10103 values between -0.10 and -0.08181 values between -0.08 and -0.06375 values between -0.06 and -0.03573 values between -0.03 and -0.01878 values between -0.01 and 0.01584 values between 0.01 and 0.04329 values between 0.04 and 0.06199 values between 0.06 and 0.08125 values between 0.08 and 0.1078 values between 0.10 and 0.1351 values between 0.13 and 0.1540 values between 0.15 and 0.1729 values between 0.17 and 0.1915 values between 0.19 and 0.2111 values between 0.21 and 0.2410 values between 0.24 and 0.263 values between 0.26 and 0.284 values between 0.28 and 0.305 values between 0.30 and 0.331 values between 0.33 and 0.352 values between 0.35 and 0.372 values between 0.37 and 0.39offnegativepositiveoften

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. @endif
  2. lu
  3. variety
  4. dạng
  5. 种类
  6. ا
  7. 恃
  8. 信息发布
  9. ApiResponse
  10. mới

Suppresses

  1. SEL
  2. issen
  3. usalem
  4. geist
  5. adden
  6. gross
  7. oha
  8. uales
  9. unded
  10. iples