© 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. 15500
    Prev
    Next
    INDEX
    Explanations

    words related to violent acts, specifically assassinations

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "assassination."

    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
    hers
    -0.75
    wm
    -0.70
    nings
    -0.69
     Consumer
    -0.68
    parent
    -0.67
    NZ
    -0.67
    umers
    -0.66
     Prev
    -0.65
     Stream
    -0.65
     bowls
    -0.64
    POSITIVE LOGITS
     assassination
    3.79
     assass
    3.46
     assassinate
    3.00
     assassinated
    2.79
     Assass
    2.77
     assassins
    2.39
     assassin
    2.37
     murder
    1.52
     ambush
    1.51
     murders
    1.43
    Activations Density 0.032%

    No Known Activations