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. 85886
    Prev
    Next
    INDEX
    Explanations

    Expressing desire to like something

    np_max-act · gemini-2.0-flash

    This neuron detects first‐person expressions of personal desire or intent, especially “I wanted to…” or “I wanted to like…” statements.

    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
    static
    -0.06
    Û
    -0.06
    _strings
    -0.06
     _(
    -0.06
     drinking
    -0.06
     Nan
    -0.06
    chrom
    -0.06
    hits
    -0.06
     엔
    -0.05
     Dj
    -0.05
    POSITIVE LOGITS
    	want
    0.07
    fail
    0.07
    .Percent
    0.07
    يع
    0.07
     distorted
    0.06
     Ή
    0.06
     výj
    0.06
     전국
    0.06
    افر
    0.06
    _scheme
    0.06
    Activations Density 0.030%

    No Known Activations