© 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. 89568
    Prev
    Next
    INDEX
    Explanations

    Models/Modelling

    np_max-act · gemini-2.0-flash

    This neuron specifically activates on the capitalized word “Model” (as in “Large Model Systems Organization”), i.e. it detects the token “Model” in proper nouns.

    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
    انی
    -0.07
    .TabControl
    -0.06
     sb
    -0.06
    _se
    -0.06
    urile
    -0.06
     गय
    -0.06
    енность
    -0.06
    етерб
    -0.06
    ARGE
    -0.06
     devote
    -0.06
    POSITIVE LOGITS
     quaint
    0.07
    .NOT
    0.07
     filesize
    0.07
    -Nov
    0.07
     домаш
    0.06
    itect
    0.06
    .addActionListener
    0.06
     model
    0.06
    propri
    0.06
    (Error
    0.06
    Activations Density 0.001%

    No Known Activations