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. 130225
    Prev
    Next
    INDEX
    Explanations

    typically

    np_max-act · gemini-2.0-flash

    The neuron flags general‐purpose procedural language—particularly adverbial cues that introduce what is usually done in a process.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    chemical compounds and their characteristics in a technical context.

    oai_token-act-pair · gpt-4o-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
     made
    -0.08
    aso
    -0.07
    resse
    -0.07
    (ad
    -0.07
    ARGE
    -0.07
    582
    -0.07
    enes
    -0.07
    =res
    -0.07
    (W
    -0.07
    agation
    -0.06
    POSITIVE LOGITS
     typically
    0.12
     Typically
    0.12
     typical
    0.11
    typically
    0.10
     Typical
    0.08
    Typ
    0.08
    pic
    0.08
     Apt
    0.08
     typ
    0.08
    elidir
    0.08
    Activations Density 0.020%

    No Known Activations