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. 129327
    Prev
    Next
    INDEX
    Explanations

    roots and powers

    np_max-act · gemini-2.0-flash

    The neuron flags the digit “1” when it appears as the numerator in fractional‐power expressions (e.g. the “1” in “1/7” or “1/4”).

    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
     مور
    -0.07
     nave
    -0.07
     hues
    -0.06
     EntityType
    -0.06
     напря
    -0.06
    .symbol
    -0.06
    FT
    -0.06
    ičky
    -0.06
    -shell
    -0.06
     validates
    -0.06
    POSITIVE LOGITS
     battled
    0.07
     topp
    0.07
     momentos
    0.07
     putting
    0.06
    LAND
    0.06
    爾
    0.06
    nému
    0.06
    licing
    0.06
    -↵
    0.06
    land
    0.06
    Activations Density 0.001%

    No Known Activations