© 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
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 25226
    Prev
    Next
    INDEX
    Explanations

    frag

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on subword tokens of “fragrance” and “perfume,” i.e. when the text refers to scents.

    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
    Meta
    -0.06
     plasma
    -0.06
    Wars
    -0.06
     Colon
    -0.06
     Schmidt
    -0.06
     days
    -0.06
     Office
    -0.06
    opt
    -0.06
     चलत
    -0.06
     Auditor
    -0.06
    POSITIVE LOGITS
     frag
    0.08
     Frag
    0.08
    	constexpr
    0.07
     :]
    0.07
     enabled
    0.07
     UICollectionViewCell
    0.06
     charm
    0.06
     avan
    0.06
    .JSONException
    0.06
    setTimeout
    0.06
    Activations Density 0.002%

    No Known Activations