© 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. 0-RES-JB
    11. 5232
    Prev
    Next
    INDEX
    Explanations

    keywords related to ease of manipulation or vulnerability

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "easily" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Deaths
    -0.67
    itness
    -0.64
    rade
    -0.64
    USS
    -0.61
    staff
    -0.59
    idelines
    -0.58
    rees
    -0.57
    ourn
    -0.56
    inas
    -0.55
    herty
    -0.55
    POSITIVE LOGITS
     easily
    3.44
     effortlessly
    2.14
     readily
    2.14
     comfortably
    1.82
     easy
    1.78
     cheaply
    1.72
     quickly
    1.69
     conveniently
    1.63
     efficiently
    1.57
     instantly
    1.56
    Activations Density 0.012%

    No Known Activations