© 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. 1-RES-JB
    11. 11588
    Prev
    Next
    INDEX
    Explanations

    terms related to political ideologies and movements

    oai_token-act-pair · gpt-3.5-turbo

    phrases associated with political themes and ideologies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    enture
    -0.62
    agraph
    -0.53
    ividual
    -0.52
    OVA
    -0.52
    imentary
    -0.51
    shall
    -0.50
     analysed
    -0.48
    otor
    -0.47
     inspected
    -0.47
    eport
    -0.46
    POSITIVE LOGITS
     ideologies
    0.66
     ideology
    0.65
     popul
    0.65
     propaganda
    0.64
     appe
    0.63
     ideologically
    0.63
     welf
    0.63
     disenfranch
    0.63
     rhetoric
    0.63
     nihil
    0.62
    Activations Density 4.880%

    No Known Activations