© 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. 1-RES-JB
    11. 21151
    Prev
    Next
    INDEX
    Explanations

    references to high levels of demand for various items or services

    oai_token-act-pair · gpt-3.5-turbo

    references to demand in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    kered
    -0.80
     Surv
    -0.69
    pleasant
    -0.64
    issance
    -0.63
     Barr
    -0.62
    ombs
    -0.62
    swick
    -0.61
    stein
    -0.61
     Aux
    -0.60
    =-=-=-=-
    -0.60
    POSITIVE LOGITS
     demand
    1.14
     Demand
    1.02
    Demand
    1.01
     demands
    0.98
    igslist
    0.94
     demanded
    0.85
    FontSize
    0.81
    demand
    0.76
     demanding
    0.75
     millenn
    0.71
    Activations Density 0.014%

    No Known Activations