© 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. 41432
    Prev
    Next
    INDEX
    Explanations

    solution, answer, review

    np_max-act · gemini-2.0-flash

    The neuron is looking for numbered list markers (tokens like “1.”, “2.”, “3.”, etc.) in the text.

    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
     BD
    -0.07
    glomer
    -0.06
     tantra
    -0.06
    占
    -0.06
     toured
    -0.06
    adora
    -0.06
    ladığı
    -0.06
     отмет
    -0.06
    imizde
    -0.06
    hits
    -0.06
    POSITIVE LOGITS
    $name
    0.08
    ंबर
    0.07
    _MAPPING
    0.07
     SECOND
    0.07
     getDate
    0.07
     Vi
    0.06
     getEmail
    0.06
    _Reset
    0.06
     onChange
    0.06
    Music
    0.06
    Activations Density 0.001%

    No Known Activations