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. 127025
    Prev
    Next
    INDEX
    Explanations

    code

    np_max-act · gemini-2.0-flash

    The neuron activates on programming‐language syntax (e.g. keywords, braces, imports, identifiers)—in other words, it detects code tokens.

    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
     counselor
    -0.07
    ologists
    -0.07
    енную
    -0.06
     probl
    -0.06
     кост
    -0.06
     cf
    -0.06
     Az
    -0.06
     salads
    -0.06
     imz
    -0.06
     yellow
    -0.06
    POSITIVE LOGITS
     []*
    0.07
     تخصصی
    0.06
     درون
    0.06
    ете
    0.06
     *</
    0.06
    issent
    0.06
    },{↵
    0.06
    俺
    0.06
    坪
    0.06
     �
    0.06
    Activations Density 0.035%

    No Known Activations