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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 17-LLAMASCOPE-RES-131K
    11. 5347
    Prev
    Next
    INDEX
    Explanations

    instances of the word "speaking" and its variations, indicating a focus on conversational or expressive phrases

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    pu
    -0.07
     Daniels
    -0.07
    kt
    -0.06
    aml
    -0.06
    amentals
    -0.06
    thon
    -0.06
    boo
    -0.06
    zk
    -0.06
    sil
    -0.06
    ingen
    -0.06
    POSITIVE LOGITS
     speaking
    0.13
    Speaking
    0.10
     Speaking
    0.09
    -speaking
    0.09
    spe
    0.09
    tae
    0.07
    umper
    0.07
     speak
    0.07
    ".$_
    0.07
    gg
    0.07
    Activations Density 0.002%

    No Known Activations