Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 85950
    Prev
    Next
    INDEX
    Explanations

    the

    np_max-act · gemini-2.0-flash

    This neuron fires on terms describing clinical‐trial outcome metrics—words like “incidence,” “occurrence,” “rate,” and related verbs (e.g. “reduce”) used to report medical endpoint frequencies.

    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
    brtc
    -0.07
     IOCTL
    -0.06
     уз
    -0.06
     freedom
    -0.06
    .Feature
    -0.06
     Mile
    -0.06
     pee
    -0.06
     refriger
    -0.06
     LPC
    -0.06
     목
    -0.06
    POSITIVE LOGITS
     contacting
    0.08
     Kiev
    0.07
    acias
    0.07
     dia
    0.06
     možnosti
    0.06
    _dates
    0.06
    _tick
    0.06
     bás
    0.06
    latable
    0.06
     Whats
    0.06
    Activations Density 0.047%

    No Known Activations