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. 1287
    Prev
    Next
    INDEX
    Explanations

    if

    np_max-act · gemini-2.0-flash

    The neuron responds to conditional clause markers that introduce hypothetical “if-then” questions.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    phrases related to evaluating safety or expressing concerns about food products.

    oai_token-act-pair · gpt-4o-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
     Afterwards
    -0.07
    και
    -0.07
    ิญญ
    -0.07
    ΜΑΤ
    -0.07
    _meter
    -0.07
    StartTime
    -0.06
     vời
    -0.06
    _LENGTH
    -0.06
    ombie
    -0.06
    .";↵
    -0.06
    POSITIVE LOGITS
    andro
    0.08
    afka
    0.08
     bara
    0.08
    那个
    0.07
     knocks
    0.06
    getNode
    0.06
    helper
    0.06
     skon
    0.06
     :
    ↵
    0.06
    -remove
    0.06
    Activations Density 0.017%

    No Known Activations