© 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. 7635
    Prev
    Next
    INDEX
    Explanations

    keywords related to security, secrets, and confidential information

    oai_token-act-pair · gpt-3.5-turbo

    references to governmental or financial regulations and secrets

    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
    kins
    -0.80
    icals
    -0.80
     Coulter
    -0.75
     Innocent
    -0.75
     Basil
    -0.71
     endowed
    -0.66
     Islam
    -0.66
    iberal
    -0.65
     Lane
    -0.65
    amacare
    -0.64
    POSITIVE LOGITS
    rets
    3.31
    RET
    2.62
    ret
    2.50
    Ret
    2.41
     RET
    1.35
    recy
    1.29
     Retro
    1.11
     retri
    1.11
     Bret
    1.09
     Tai
    1.06
    Activations Density 0.032%

    No Known Activations