© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Qwen3-4B
    3. 24-TRANSCODER-HP
    4. 150274
    Prev
    Next
    INDEX
    Explanations

    exact, precise

    np_max-act-logits · gemini-2.0-flash

    transitional or contrasting phrases that introduce exceptions or counterpoints in academic/scientific writing (words like "but", "however", "yet", "while" that signal exceptions to the preceding statement).

    oai_token-act-pair · deepseek-v3Triggered by @finnwit
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_24.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.24.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    enity
    -0.27
    ä¸įè§ģäºĨ
    -0.26
    ió
    -0.25
    edom
    -0.24
     Nolan
    -0.24
    ä¸į管æĺ¯
    -0.23
    NDAR
    -0.23
     Checked
    -0.23
    rray
    -0.23
     appellant
    -0.23
    POSITIVE LOGITS
     exact
    1.14
    çļĦåħ·ä½ĵ
    0.96
    åħ·ä½ĵçļĦ
    0.94
    exact
    0.94
    åħ·ä½ĵ
    0.93
     Exact
    0.90
    Exact
    0.88
     precise
    0.87
    _exact
    0.83
     exactly
    0.81
    Activations Density 1.091%

    No Known Activations