© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 39108
    Prev
    Next
    INDEX
    Explanations

    unfamiliar

    np_max-act · gemini-2.0-flash

    The neuron is primarily activated by numerical tokens (digits, numbers, or numeric expressions).

    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
     earned
    -0.08
     Joanna
    -0.07
     memcpy
    -0.07
     Tr
    -0.06
     chứa
    -0.06
     utilizando
    -0.06
    :int
    -0.06
     debugger
    -0.06
    δικ
    -0.06
     уб
    -0.06
    POSITIVE LOGITS
     umíst
    0.07
     onFinish
    0.06
     �
    0.06
    (Map
    0.06
    ĩnh
    0.06
    '])[
    0.06
     #:
    0.06
    ural
    0.06
    ')).
    0.06
    _scroll
    0.06
    Activations Density 0.012%

    No Known Activations