Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 93093
    Prev
    Next
    INDEX
    Explanations

    hyphen

    np_max-act · gemini-2.0-flash

    The neuron activates on chemical element symbols and bond/angle notation (e.g. “Te,” “C–Te–C,” “C---B1”), effectively spotting element names in bond descriptions.

    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
    ansa
    -0.07
    桥
    -0.07
    pants
    -0.07
    endir
    -0.06
    Incoming
    -0.06
     Chapel
    -0.06
    arnation
    -0.06
    .ObjectMapper
    -0.06
    geh
    -0.06
     taşın
    -0.06
    POSITIVE LOGITS
     GIF
    0.06
    (turn
    0.06
    .Horizontal
    0.06
     jadx
    0.06
     φ
    0.06
    _likelihood
    0.06
     showcasing
    0.06
     enforcing
    0.06
    hdr
    0.06
     Dis
    0.06
    Activations Density 0.013%

    No Known Activations