© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 3-RES-JB
    11. 9210
    Prev
    Next
    INDEX
    Explanations

    mentions of reports

    oai_token-act-pair · gpt-3.5-turbo

    mentions of reports, particularly in a formal or analytical context

    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
    creen
    -0.99
    east
    -0.75
    cffff
    -0.74
     Klux
    -0.70
    conservancy
    -0.69
    othing
    -0.66
    ntil
    -0.64
    venge
    -0.62
    vous
    -0.61
    姫
    -0.61
    POSITIVE LOGITS
    report
    0.84
    books
    0.81
    Crash
    0.78
     report
    0.75
     reports
    0.74
    imony
    0.71
     synopsis
    0.71
    ounces
    0.70
    Reporting
    0.69
    Versions
    0.68
    Activations Density 0.042%

    No Known Activations