© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 32-TRANSCODER-HP
    7. 41032
    Prev
    Next
    INDEX
    Explanations

    say "not" words

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
     rather
    -0.49
    rather
    -0.45
    èĢĮä¸įæĺ¯
    -0.39
     Rather
    -0.38
     plutôt
    -0.36
    èĢĮéĿŀ
    -0.36
    Rather
    -0.35
     chứ
    -0.33
    ä¸įæ¸ħ
    -0.31
    èĢĮä¸į
    -0.29
    POSITIVE LOGITS
     net
    0.41
    -N
    0.39
     nu
    0.39
     ot
    0.38
    -na
    0.38
     nat
    0.37
     NU
    0.37
     nid
    0.37
    _nf
    0.36
     NI
    0.35
    Activations Density 0.036%

    No Known Activations