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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 84678
    Prev
    Next
    INDEX
    Explanations

    and

    np_max-act · gemini-2.0-flash

    This neuron responds to alphanumeric technical identifiers and acronyms (e.g. gene/protein names, assay codes, chemical formulas).

    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
     giving
    -0.07
    .after
    -0.06
     оч
    -0.06
    	cc
    -0.06
    take
    -0.06
    leck
    -0.06
     Patton
    -0.06
    .interval
    -0.06
     tầm
    -0.06
     front
    -0.06
    POSITIVE LOGITS
     και
    0.08
     and
    0.08
     tane
    0.07
     và
    0.07
     or
    0.07
     AND
    0.07
    覧
    0.06
     Ox
    0.06
     &&
    0.06
     форма
    0.06
    Activations Density 0.214%

    No Known Activations