© 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 (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 41647
    Prev
    Next
    INDEX
    Explanations

    Numbers/letters for communication

    np_max-act · gemini-2.0-flash

    This neuron activates on numeric tokens and digit sequences (e.g., numbers, time stamps, phone‐style digits).

    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
    _wr
    -0.06
     Slide
    -0.06
    Interview
    -0.06
    Equivalent
    -0.06
    ์ล
    -0.06
     Emp
    -0.06
    �
    -0.06
     Erg
    -0.06
    PUT
    -0.06
     武
    -0.05
    POSITIVE LOGITS
    ossa
    0.07
     humour
    0.07
     grams
    0.07
    ículo
    0.06
     Netz
    0.06
     guessing
    0.06
    حيح
    0.06
    iline
    0.06
    _fetch
    0.06
     prospect
    0.06
    Activations Density 0.015%

    No Known Activations