© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 129583
    Prev
    Next
    INDEX
    Explanations

    to

    np_max-act · gemini-2.0-flash

    This neuron fires on “how to” instructions or other procedural cues, identifying phrases that introduce steps or guidance (e.g. “how to apply,” “how to return”).

    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
    instances
    -0.06
     Dor
    -0.06
    tables
    -0.06
     '\\
    -0.06
    Much
    -0.06
     центр
    -0.06
    .Mapping
    -0.06
    Bot
    -0.05
    90
    -0.05
     kol
    -0.05
    POSITIVE LOGITS
     renewable
    0.07
     Česk
    0.07
     ModelRenderer
    0.07
     ssid
    0.07
     Arabia
    0.07
    /count
    0.07
    _PRIV
    0.07
    �
    0.06
     investigación
    0.06
     HACK
    0.06
    Activations Density 0.036%

    No Known Activations