Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 14422
    Prev
    Next
    INDEX
    Explanations

    words related to pedestrian walkways and infrastructure

    oai_token-act-pair · gpt-3.5-turbo

    variations of the word "rian" or related suffixes

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     disp
    -0.72
     Powder
    -0.71
     ups
    -0.70
     wavelength
    -0.65
     extract
    -0.65
     Fortune
    -0.65
     pow
    -0.65
     immun
    -0.63
     envelop
    -0.61
     whit
    -0.61
    POSITIVE LOGITS
    rian
    4.78
    rians
    3.74
    ria
    2.32
    rien
    1.59
    rius
    1.53
    riad
    1.40
    rium
    1.37
    alian
    1.28
    gian
    1.26
    rious
    1.23
    Activations Density 0.009%

    No Known Activations