© 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. 12245
    Prev
    Next
    INDEX
    Explanations

    words related to specific locations or place names

    oai_token-act-pair · gpt-3.5-turbo

    mentions of locations or places, particularly those with the suffix 'ville'

    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
    Wan
    -0.80
    ittal
    -0.74
    asel
    -0.67
    articles
    -0.66
    senal
    -0.66
    wcsstore
    -0.65
    unes
    -0.65
    BOOK
    -0.64
     paras
    -0.63
    ighth
    -0.63
    POSITIVE LOGITS
    ville
    1.14
    dale
    0.80
     Heights
    0.79
    iculture
    0.78
    gomery
    0.78
     County
    0.76
    ites
    0.75
    ite
    0.75
     Correctional
    0.75
    sburgh
    0.74
    Activations Density 0.014%

    No Known Activations