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. 130647
    Prev
    Next
    INDEX
    Explanations

    Start of text

    np_max-act · gemini-2.0-flash

    The neuron activates on high-level section header and title words (like “Applications,” “Properties,” “Introduction,” etc.) marking the starts of major article sections.

    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
    олева
    -0.06
    cmds
    -0.06
    uida
    -0.06
    ffee
    -0.06
     안
    -0.06
    _stride
    -0.06
    userService
    -0.06
     Phase
    -0.06
     graphs
    -0.06
     trie
    -0.06
    POSITIVE LOGITS
    becue
    0.07
    (nr
    0.06
    -native
    0.06
     Queensland
    0.06
     OTP
    0.06
     blessing
    0.06
     mac
    0.06
     penchant
    0.06
     options
    0.06
    .selector
    0.06
    Activations Density 0.040%

    No Known Activations