© 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. 8-RES-JB
    11. 988
    Prev
    Next
    INDEX
    Explanations

    sequences of words separated by strange characters ("Ċ")

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of numerical values or sequences in the text

    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
     payroll
    -0.78
     toll
    -0.72
     civilian
    -0.67
     caus
    -0.66
     favor
    -0.65
     demos
    -0.64
     FML
    -0.64
     iod
    -0.64
     therm
    -0.63
     rye
    -0.63
    POSITIVE LOGITS
    Speaking
    1.18
    British
    1.14
    Australian
    1.11
    Despite
    1.10
    Hum
    1.08
    Former
    1.08
    Actor
    1.07
    Britain
    1.07
    Parents
    1.07
    However
    1.06
    Activations Density 0.176%

    No Known Activations