© 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. 3-RES-JB
    11. 12969
    Prev
    Next
    INDEX
    Explanations

    words related to corruption and scandal

    oai_token-act-pair · gpt-3.5-turbo

    the word "rob" and its variations, indicating a focus on concepts related to theft or robbery

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    hips
    -0.75
     Leilan
    -0.68
    alam
    -0.66
    DN
    -0.65
     largeDownload
    -0.63
    åħī
    -0.62
     Malays
    -0.61
     Sakuya
    -0.61
     FORM
    -0.61
     LINE
    -0.60
    POSITIVE LOGITS
    bing
    1.07
    atically
    1.04
    ooth
    0.99
    acter
    0.99
    bers
    0.96
    aton
    0.93
    otech
    0.90
    rob
    0.89
    axter
    0.89
    rities
    0.89
    Activations Density 0.017%

    No Known Activations