© 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. 8-RES-JB
    11. 24118
    Prev
    Next
    INDEX
    Explanations

    terms related to predictions or possibilities

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating inability or failure

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    gypt
    -0.67
    OTOS
    -0.64
     periphery
    -0.58
     assembled
    -0.57
     constrained
    -0.56
     bonded
    -0.56
     guiding
    -0.56
     compr
    -0.55
     guided
    -0.54
     populated
    -0.54
    POSITIVE LOGITS
    't
    1.57
    cest
    1.14
    itive
    1.08
    now
    0.97
    ipeg
    0.95
    nings
    0.92
    kered
    0.92
    throp
    0.91
    cing
    0.89
    iors
    0.88
    Activations Density 0.034%

    No Known Activations