© 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. 722
    Prev
    Next
    INDEX
    Explanations

    references to a specific brand or company, specifically "Bethesda" and some related mentions of "Tomb Raider" and "Anita" individuals

    oai_token-act-pair · gpt-3.5-turbo

    references to the video game company Bethesda and the "Tomb Raider" franchise

    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
    psy
    -0.69
    FLAG
    -0.68
    calling
    -0.68
    asin
    -0.67
    Assad
    -0.67
    nda
    -0.66
    inel
    -0.65
    KK
    -0.65
    isting
    -0.64
    most
    -0.64
    POSITIVE LOGITS
    hesda
    1.56
     Bethesda
    1.21
     Soft
    0.89
    halla
    0.84
     Oblivion
    0.78
    reath
    0.76
     Blizzard
    0.74
    sburg
    0.73
    romeda
    0.73
    sonian
    0.73
    Activations Density 0.010%

    No Known Activations