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. 17131
    Prev
    Next
    INDEX
    Explanations

    words related to furniture, particularly desks

    oai_token-act-pair · gpt-3.5-turbo

    mentions of desks and related furniture

    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
    na
    -0.76
    ĨĴ
    -0.73
     Boyle
    -0.72
    ás
    -0.68
    aila
    -0.65
    á
    -0.62
    asher
    -0.61
    annot
    -0.60
    onga
    -0.60
    ÏĦ
    -0.59
    POSITIVE LOGITS
     desk
    4.07
     desks
    3.23
     Desk
    2.61
    Desk
    1.98
     desktop
    1.46
     office
    1.32
     sofa
    1.22
     typew
    1.20
     table
    1.19
    office
    1.19
    Activations Density 0.010%

    No Known Activations