Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 3741
    Prev
    Next
    INDEX
    Explanations

    phrases related to an online game or virtual environment

    oai_token-act-pair · gpt-3.5-turbo

    the word "verse" and its variants, indicating a focus on poetic or lyrical content

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     ice
    -0.72
     Acer
    -0.70
     espresso
    -0.68
     Jelly
    -0.62
     award
    -0.61
     Candy
    -0.60
     hamb
    -0.60
     Gren
    -0.60
     Beg
    -0.59
     videot
    -0.59
    POSITIVE LOGITS
    verse
    5.04
    vers
    1.62
    VERS
    1.57
    versely
    1.46
    versions
    1.34
    vert
    1.18
    verted
    1.17
    version
    1.15
    Ver
    0.99
    verting
    0.95
    Activations Density 0.016%

    No Known Activations