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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 110740
    Prev
    Next
    INDEX
    Explanations

    Shakespearean plays

    np_max-act · gemini-2.0-flash

    This neuron gives stronger positive activations to rarer, less frequent content‐carrying tokens and near zero to common or structural tokens.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Covenant
    -0.07
    makt
    -0.07
     tvor
    -0.06
     ederek
    -0.06
     модель
    -0.06
     Harbour
    -0.06
    .Remove
    -0.06
    _PIPELINE
    -0.06
     сразу
    -0.06
     Pip
    -0.06
    POSITIVE LOGITS
    >'.
    0.07
    ."</
    0.07
     refs
    0.06
    атель
    0.06
     İt
    0.06
    aying
    0.06
     древ
    0.06
     Resistance
    0.06
     paying
    0.06
    neum
    0.06
    Activations Density 0.010%

    No Known Activations