© 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. 1672
    Prev
    Next
    INDEX
    Explanations

    proper nouns commonly associated with news organizations

    oai_token-act-pair · gpt-3.5-turbo

    mentions of specific geographical locations, primarily "San" and "New York"

    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
     Elect
    -0.77
     Eliot
    -0.74
     Lamp
    -0.71
     Wheel
    -0.68
     Nixon
    -0.67
     Crow
    -0.66
     notes
    -0.66
     Chart
    -0.65
     Poll
    -0.65
     Weld
    -0.65
    POSITIVE LOGITS
    SAN
    4.06
     SAN
    2.59
    AMI
    1.60
     YORK
    1.45
     SAP
    1.41
    LOS
    1.36
    LAN
    1.29
    WASHINGTON
    1.28
    ANA
    1.23
    STON
    1.20
    Activations Density 0.020%

    No Known Activations