© 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 EvalsExportsinterp-engineNEWAPI 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. 4-RES-JB
    11. 12816
    Prev
    Next
    INDEX
    Explanations

    phrases involving actions or events that have a negative or controversial connotation

    oai_token-act-pair · gpt-3.5-turbo

    negative expressions or concepts related to being off-duty

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

    No Comments

    Negative Logits
    ulhu
    -0.92
     Dickinson
    -0.77
     reviewer
    -0.70
     Vid
    -0.70
     Compass
    -0.69
     Burr
    -0.68
     accuser
    -0.67
     Cassidy
    -0.66
     Zar
    -0.66
     EntityItem
    -0.66
    POSITIVE LOGITS
    sized
    1.28
    powered
    1.08
    tun
    1.07
    balance
    1.05
    aligned
    1.05
    level
    1.04
    earth
    1.04
    grade
    1.04
    performing
    1.03
    bred
    1.03
    Activations Density 0.119%

    No Known Activations