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. 129446
    Prev
    Next
    INDEX
    Explanations

    leave

    np_max-act · gemini-2.0-flash

    descriptions of organized spaces or storage for children’s toys.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron activates on phrasing that suggests leaving something “as is,” e.g. “leave this … ‘as is.’”

    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
     ]);
    -0.07
    ните
    -0.07
     Companion
    -0.07
    �
    -0.07
    Management
    -0.07
     intimacy
    -0.07
     Circuit
    -0.07
    ें
    -0.06
    Probability
    -0.06
     nanoparticles
    -0.06
    POSITIVE LOGITS
     Leave
    0.07
     gv
    0.06
     leave
    0.06
     ogl
    0.06
     留
    0.06
     лим
    0.06
    VERS
    0.06
    vers
    0.06
    uč
    0.06
     Prints
    0.06
    Activations Density 0.016%

    No Known Activations