© 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. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 6-LLAMASCOPE-RES-32K
    11. 25495
    Prev
    Next
    INDEX
    Explanations

    phrases emphasizing actions or directives involving "to", often related to goals, intentions, or purposes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.1-8B @ 6-llamascope-res-32k
    Configuration
    fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L6R-8x
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    Features
    32,768
    Data Type
    bfloat16
    Hook Name
    blocks.6.hook_resid_post
    Hook Layer
    6
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    cerebras/SlimPajama-627B
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Lowe
    -0.16
     tele
    -0.15
    itori
    -0.15
    iosa
    -0.15
    edi
    -0.14
     nouve
    -0.14
    thest
    -0.14
     precisely
    -0.14
    uler
    -0.13
    jaw
    -0.13
    POSITIVE LOGITS
    uptools
    0.15
    enda
    0.15
    aid
    0.14
    Forgery
    0.14
    ãĥĥãĥģ
    0.14
    umor
    0.13
    ylko
    0.13
    커ìĬ¤
    0.13
    \Doctrine
    0.13
     bear
    0.13
    Activations Density 0.030%

    No Known Activations