Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 85974
    Prev
    Next
    INDEX
    Explanations

    top layer

    np_max-act · gemini-2.0-flash

    The neuron consistently activates on phrases referring to the “top” or “uppermost” layer or surface (e.g. “top layer,” “surface,” “top layers”).

    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
     рассчит
    -0.07
     Gloria
    -0.07
     Otto
    -0.07
     Ramos
    -0.06
    ests
    -0.06
     Григор
    -0.06
    iu
    -0.06
     transmitted
    -0.06
    .Zip
    -0.06
     Sala
    -0.06
    POSITIVE LOGITS
    elay
    0.07
    iostream
    0.07
     TLabel
    0.06
    -state
    0.06
    -shopping
    0.06
    خب
    0.06
     nam
    0.06
     Değ
    0.06
    .pkg
    0.06
    -ignore
    0.06
    Activations Density 0.027%

    No Known Activations