© 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. 30-TRANSCODER-HP
    7. 97058
    Prev
    Next
    INDEX
    Explanations

    was/were

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    æĹ©çĤ¹
    -0.40
     early
    -0.35
    æĹ©æľŁ
    -0.28
    å°½æĹ©
    -0.28
    early
    -0.27
    -publish
    -0.26
    æĹ©
    -0.25
    preter
    -0.25
    _pen
    -0.25
    æĭĺ
    -0.25
    POSITIVE LOGITS
     around
    0.41
     Around
    0.34
    around
    0.31
     at
    0.31
    atus
    0.30
    rom
    0.28
    Around
    0.27
     Harness
    0.27
    At
    0.27
    -around
    0.27
    Activations Density 0.886%

    No Known Activations