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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 17537
    Prev
    Next
    INDEX
    Explanations

    references to valuable items or things that are highly prized

    oai_token-act-pair · gpt-3.5-turbo

    references to valuable items or collectibles

    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
    okin
    -0.72
    ells
    -0.70
    colo
    -0.67
    ellar
    -0.66
    appa
    -0.66
    abad
    -0.66
    imov
    -0.65
    arel
    -0.65
    SPA
    -0.65
    slow
    -0.64
    POSITIVE LOGITS
     trove
    1.44
     treasure
    0.91
     treasures
    0.87
    vier
    0.83
     riches
    0.82
     chests
    0.81
    abilia
    0.81
    collection
    0.77
     Treasure
    0.75
    asures
    0.72
    Activations Density 0.010%

    No Known Activations