© 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. 11-RES-JB
    11. 10297
    Prev
    Next
    INDEX
    Explanations

    informational terms such as "information", "evidence", "data", and "address"

    oai_token-act-pair · gpt-3.5-turbo

    keywords related to information and documentation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     sqor
    -0.70
    +.
    -0.70
    ãĤ´ãĥ³
    -0.69
    ãĤ¡
    -0.69
    poon
    -0.63
    bsp
    -0.63
    ullivan
    -0.61
    %.
    -0.61
    ECA
    -0.61
    yet
    -0.60
    POSITIVE LOGITS
     varies
    1.30
     consists
    1.23
     differs
    1.19
     derives
    1.18
     corresponds
    1.16
     depends
    1.16
     represents
    1.14
     belongs
    1.14
     reflects
    1.13
     is
    1.12
    Activations Density 0.434%

    No Known Activations