© 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. 8-RES-JB
    11. 14864
    Prev
    Next
    INDEX
    Explanations

    words related to significant negative impact or outcome

    oai_token-act-pair · gpt-3.5-turbo

    descriptors of severe damage or destruction

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    cius
    -0.85
    skirts
    -0.78
    cipl
    -0.74
    Frames
    -0.73
    uni
    -0.71
    qi
    -0.71
    haps
    -0.70
    ploma
    -0.70
    rist
    -0.70
    etch
    -0.69
    POSITIVE LOGITS
     devastation
    1.00
     blow
    0.96
     consequences
    0.94
     impacts
    0.94
    ly
    0.91
     blows
    0.90
     havoc
    0.90
     damage
    0.88
     disasters
    0.87
     aftermath
    0.86
    Activations Density 0.064%

    No Known Activations