© 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. 11-RES-JB
    11. 916
    Prev
    Next
    INDEX
    Explanations

    words related to spatial and temporal relationships such as 'in', 'from', 'on', and 'for'

    oai_token-act-pair · gpt-3.5-turbo

    prepositions and their usage in context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     EVERY
    -0.71
     everywhere
    -0.61
    ¶
    -0.54
     IMAGES
    -0.52
     cried
    -0.52
    hops
    -0.51
     screamed
    -0.51
    everyone
    -0.51
    certain
    -0.50
    ordinary
    -0.50
    POSITIVE LOGITS
     whom
    1.06
     which
    0.81
     whose
    0.79
    which
    0.77
    whose
    0.76
     lieu
    0.73
     whence
    0.72
     respectively
    0.62
     behalf
    0.61
     recommendation
    0.60
    Activations Density 0.369%

    No Known Activations