© 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. 4-RES-JB
    11. 4512
    Prev
    Next
    INDEX
    Explanations

    references to a specific person or entity named "Rozz"

    oai_token-act-pair · gpt-3.5-turbo

    repeated references to the name "Ro" or "Razz" in various contexts

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

    No Comments

    Negative Logits
     Polar
    -0.73
     lapse
    -0.72
     fitness
    -0.70
     Martial
    -0.68
    vict
    -0.68
     Conrad
    -0.67
     foremost
    -0.65
     Scots
    -0.65
     unthinkable
    -0.64
     conditioning
    -0.64
    POSITIVE LOGITS
    zz
    1.40
    arella
    1.12
    ZZ
    1.07
    ucc
    0.98
    azz
    0.98
    abba
    0.96
    arro
    0.94
    hou
    0.93
    zzle
    0.93
    edd
    0.91
    Activations Density 0.014%

    No Known Activations