© 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. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 32916
    Prev
    Next
    INDEX
    Explanations

    instances of the word "it" and variations of the verb "to be"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ufs
    -0.14
    nestjs
    -0.14
     CircularProgress
    -0.14
    elda
    -0.14
    odus
    -0.14
    /browse
    -0.14
    osh
    -0.14
    æķij
    -0.14
    ÏĢει
    -0.14
    esson
    -0.13
    POSITIVE LOGITS
     would
    0.26
     became
    0.23
     wouldn
    0.22
     Would
    0.20
     wasn
    0.20
     was
    0.19
    Would
    0.18
    would
    0.18
     zou
    0.18
     seemed
    0.17
    Activations Density 0.027%

    No Known Activations