© 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. 7301
    Prev
    Next
    INDEX
    Explanations

    proper nouns of people, places, and organizations, with a focus on individuals named Ernest and references to Algeria

    oai_token-act-pair · gpt-3.5-turbo

    references to specific individuals and locations associated with Algerian history or culture

    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
    peria
    -0.85
    secution
    -0.76
    inki
    -0.76
    pered
    -0.76
    zyme
    -0.74
    pering
    -0.71
    ghai
    -0.71
    ku
    -0.70
    orks
    -0.69
    oval
    -0.69
    POSITIVE LOGITS
    aneously
    0.77
    aneous
    0.74
    orical
    0.71
    antly
    0.70
     Ernest
    0.69
     fatig
    0.66
     dysph
    0.66
    rained
    0.65
     Alger
    0.65
    ellig
    0.65
    Activations Density 0.025%

    No Known Activations