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. 3-RES-JB
    6. 6769
    Prev
    Next
    INDEX
    Explanations

    proper nouns with initials or abbreviations

    oai_token-act-pair · gpt-3.5-turbo

    capital letters, particularly of names or titles

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ModLoader
    -0.81
    dylib
    -0.75
    Topics
    -0.71
    FACE
    -0.68
    culosis
    -0.68
    Relations
    -0.67
    20439
    -0.67
    Contents
    -0.66
    duino
    -0.65
    âĶĢâĶĢ
    -0.64
    POSITIVE LOGITS
    itte
    0.78
    ahn
    0.74
    iner
    0.73
    uty
    0.70
    ze
    0.70
    oust
    0.69
    oner
    0.69
    utsch
    0.67
    erman
    0.67
    ham
    0.66
    Activations Density 0.257%

    No Known Activations