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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 85013
    Prev
    Next
    INDEX
    Explanations

    code or API keys

    np_max-act · gemini-2.0-flash

    This neuron detects occurrences of place‐name fragments beginning with “Dh” (e.g. Dharamshala, Dhauladhar).

    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
    2
    -0.08
     напрям
    -0.08
    นต
    -0.07
     کام
    -0.07
     Flor
    -0.07
    .getStringExtra
    -0.07
    ับต
    -0.07
    ))*
    -0.07
    ilebilir
    -0.06
    ivid
    -0.06
    POSITIVE LOGITS
    h
    0.16
    H
    0.15
    CH
    0.14
    ch
    0.13
    th
    0.13
    <h
    0.12
     DH
    0.12
     gh
    0.12
    sh
    0.12
    TH
    0.12
    Activations Density 0.834%

    No Known Activations