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. 4-RES-JB
    6. 12683
    Prev
    Next
    INDEX
    Explanations

    words related to the capability or potential for certain actions or modifications

    oai_token-act-pair · gpt-3.5-turbo

    words related to the concept of being capable or functional

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 4-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    4
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    itect
    -0.79
    erva
    -0.76
    oult
    -0.69
    udeb
    -0.69
    ournal
    -0.67
    resa
    -0.66
    arte
    -0.65
    ffield
    -0.65
    atcher
    -0.64
    acha
    -0.64
    POSITIVE LOGITS
    MpServer
    0.75
    nesses
    0.71
    Ptr
    0.69
    Redditor
    0.68
    pedia
    0.68
    Characters
    0.67
    ible
    0.67
    lihood
    0.67
     idiots
    0.66
    lycer
    0.66
    Activations Density 0.028%

    No Known Activations