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

    Neuronpedia

    Jacobian LensNEW
    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. 10-RES-JB
    11. 6020
    Prev
    Next
    INDEX
    Explanations

    mentions of specific locations or historical/figurative figures

    oai_token-act-pair · gpt-3.5-turbo

    repeated instances of the same root words or similar phonetic patterns

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    wives
    -0.72
    ĸļ
    -0.71
    lance
    -0.66
    abilia
    -0.63
     hus
    -0.63
    ciating
    -0.58
    ascript
    -0.58
    lishing
    -0.56
     Doodle
    -0.55
    earable
    -0.55
    POSITIVE LOGITS
    lehem
    0.85
    inas
    0.85
    acci
    0.80
    aux
    0.78
    aire
    0.71
    iar
    0.70
    illet
    0.70
    ainment
    0.69
     Beaut
    0.68
    nice
    0.67
    Activations Density 0.165%

    No Known Activations