© 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. 11-RES-JB
    11. 20959
    Prev
    Next
    INDEX
    Explanations

    words related to fatal or deadly events

    oai_token-act-pair · gpt-3.5-turbo

    terms associated with lethal or fatal incidents

    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
    erous
    -0.97
    orthy
    -0.81
    yi
    -0.80
    yang
    -0.79
    wright
    -0.75
    ourced
    -0.72
    unda
    -0.72
    Newsletter
    -0.67
    here
    -0.67
    nai
    -0.67
    POSITIVE LOGITS
     flaw
    1.05
     overdoses
    0.99
     overdose
    0.95
     gunshot
    0.95
     dose
    0.94
     injection
    0.94
     blow
    0.91
     consequences
    0.90
     combination
    0.89
     shootings
    0.88
    Activations Density 0.038%

    No Known Activations