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. 90066
    Prev
    Next
    INDEX
    Explanations

    name

    np_max-act · gemini-2.0-flash

    This neuron selectively activates on instructions asking for “tool names,” i.e. phrases specifying to output only the names of tools.

    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
     parameters
    -0.06
     pipes
    -0.06
    illage
    -0.06
    .currentUser
    -0.06
    -share
    -0.06
     uk
    -0.06
    .circular
    -0.06
     стандарт
    -0.06
     ве
    -0.06
     saldırı
    -0.06
    POSITIVE LOGITS
    ίνει
    0.07
    acle
    0.07
     affair
    0.07
    дии
    0.06
     motto
    0.06
    .cljs
    0.06
     bekl
    0.06
     Ult
    0.06
     اختل
    0.06
     Tibetan
    0.06
    Activations Density 0.012%

    No Known Activations