© 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. 58034
    Prev
    Next
    INDEX
    Explanations

    Scientific texts

    np_max-act · gemini-2.0-flash

    The neuron is most strongly activated by short, high‐frequency function words and small common tokens such as “a,” “of,” and “from.”

    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
    -only
    -0.06
    Maybe
    -0.06
     Shank
    -0.06
    cards
    -0.06
    Night
    -0.05
     překvap
    -0.05
     comparison
    -0.05
     Subway
    -0.05
     Cylinder
    -0.05
     Memo
    -0.05
    POSITIVE LOGITS
    690
    0.08
    vlc
    0.07
     отмеч
    0.07
    tal
    0.06
    imler
    0.06
    кус
    0.06
     Hosp
    0.06
    속
    0.06
    &q
    0.06
    .writeln
    0.06
    Activations Density 0.088%

    No Known Activations