© 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. 39106
    Prev
    Next
    INDEX
    Explanations

    metaph

    np_max-act · gemini-2.0-flash

    This neuron selectively activates on technical philosophy terms—especially names of subfields like “metaphysics,” “ontology,” “epistemology,” and similar headings.

    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
     shameful
    -0.06
    .board
    -0.06
     simp
    -0.06
    .Pro
    -0.06
    .cols
    -0.06
     disgrace
    -0.06
     interchange
    -0.06
     hooked
    -0.06
    .setStroke
    -0.06
     народ
    -0.06
    POSITIVE LOGITS
    (PATH
    0.07
     condos
    0.07
    mental
    0.07
    墨
    0.06
     Б
    0.06
    _UUID
    0.06
    AGED
    0.06
    住宅
    0.06
    atrib
    0.06
    าคา
    0.06
    Activations Density 0.006%

    No Known Activations