© 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 (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 788
    Prev
    Next
    INDEX
    Explanations

    finding solutions/methods

    np_max-act · gemini-2.0-flash

    This neuron detects the relative‐clause prepositional pattern “in which to,” i.e. phrases like “in which to study.”

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

    phrases expressing searching or needing to find/locate something, often invoking an unspecified place or solution.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
    /grpc
    -0.07
    _routes
    -0.07
    ائم
    -0.07
    -born
    -0.07
    ù
    -0.06
    ором
    -0.06
     handjob
    -0.06
    -session
    -0.06
     folded
    -0.06
    ژن
    -0.06
    POSITIVE LOGITS
     Birthday
    0.07
    Cel
    0.07
     Sparks
    0.06
     Cel
    0.06
     teammates
    0.06
    kil
    0.06
    关键
    0.06
    emma
    0.06
     平方
    0.06
     Alps
    0.06
    Activations Density 0.262%

    No Known Activations