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. 85976
    Prev
    Next
    INDEX
    Explanations

    is

    np_max-act · gemini-2.0-flash

    The neuron specializes in detecting numeric result tokens, i.e. floating‐point values output by the calculations.

    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
    /rem
    -0.07
    gent
    -0.06
    urd
    -0.06
    ullet
    -0.06
     کودک
    -0.06
     arrests
    -0.06
     wardrobe
    -0.06
     порушення
    -0.06
    aber
    -0.06
     अब
    -0.06
    POSITIVE LOGITS
    (proxy
    0.07
     useful
    0.06
    Accessible
    0.06
     acceler
    0.06
     Например
    0.06
    "?↵↵
    0.06
     approx
    0.06
    	send
    0.06
     spoke
    0.06
    -made
    0.06
    Activations Density 0.004%

    No Known Activations