© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 7050
    Prev
    Next
    INDEX
    Explanations

    references to geographical locations with the abbreviation "NH" along with some names like "Jeanne"

    oai_token-act-pair · gpt-3.5-turbo

    mentions of specific geographical locations

    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
    cial
    -0.83
    eering
    -0.75
    module
    -0.73
    deck
    -0.70
    notations
    -0.70
    ãĥī
    -0.68
    ãĥ´ãĤ¡
    -0.67
    aic
    -0.67
    minster
    -0.67
    initialized
    -0.65
    POSITIVE LOGITS
    NH
    1.15
     NH
    0.86
    urst
    0.82
    IGH
    0.79
    UFC
    0.78
    eree
    0.78
    MH
    0.77
    ADRA
    0.76
    JD
    0.75
    ARB
    0.75
    Activations Density 0.008%

    No Known Activations