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. 104283
    Prev
    Next
    INDEX
    Explanations

    Numerical references

    np_max-act · gemini-2.0-flash

    This neuron never fires on any ordinary text tokens—i.e. it’s effectively “dead” (or tuned to a feature not present in these excerpts).

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

    phrases related to harmful bacteria and their effects on human health.

    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
     archive
    -0.07
    z
    -0.07
    Session
    -0.07
     elements
    -0.07
     AE
    -0.07
     Funk
    -0.06
    archive
    -0.06
     folded
    -0.06
     eligible
    -0.06
     objectId
    -0.06
    POSITIVE LOGITS
     νο
    0.07
    <const
    0.07
     eventData
    0.06
    ��
    0.06
    llu
    0.06
     asoci
    0.06
    =get
    0.06
    intptr
    0.06
    \data
    0.06
     روم
    0.06
    Activations Density 0.044%

    No Known Activations