© 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. 6148
    Prev
    Next
    INDEX
    Explanations

    Names in quotes

    np_max-act · gemini-2.0-flash

    This neuron detects speaker attribution phrases—names (often coaches) paired with reporting verbs (e.g. “said,” “said Nolan,” “coach X said”).

    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
    燕
    -0.06
    752
    -0.06
     เซ
    -0.06
    242
    -0.06
    秦
    -0.05
     spline
    -0.05
    setw
    -0.05
     relics
    -0.05
     resultMap
    -0.05
    Longitude
    -0.05
    POSITIVE LOGITS
     falling
    0.07
     admitting
    0.07
    _n
    0.06
     Mister
    0.06
     lieutenant
    0.06
    ('--
    0.06
    >/
    0.06
    (Profile
    0.06
    /status
    0.06
     заним
    0.06
    Activations Density 0.016%

    No Known Activations