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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 89660
    Prev
    Next
    INDEX
    Explanations

    Categories

    np_max-act · gemini-2.0-flash

    This neuron fires on proper‐noun tokens appearing in category or metadata lines—particularly geographic names (e.g. Connecticut, Hartford) and institutional bodies (e.g. United States House).

    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
     geography
    -0.07
    SORT
    -0.07
     PLACE
    -0.07
    UCE
    -0.07
     Cheer
    -0.07
     Sys
    -0.07
     Ecc
    -0.06
    _DIALOG
    -0.06
    —is
    -0.06
    antics
    -0.06
    POSITIVE LOGITS
     Eq
    0.06
    甚至
    0.06
     Vậy
    0.06
     небольш
    0.06
    ="-
    0.06
    .be
    0.06
    })↵↵↵
    0.06
    Neutral
    0.06
    avigate
    0.06
    tracer
    0.05
    Activations Density 0.008%

    No Known Activations