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

    Neuronpedia

    Jacobian LensNEW
    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. 8-RES-JB
    11. 10357
    Prev
    Next
    INDEX
    Explanations

    words related to light and sound

    oai_token-act-pair · gpt-3.5-turbo

    references to lights and illumination in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Flavoring
    -0.76
    Form
    -0.75
    obos
    -0.71
     Wrest
    -0.69
     Arche
    -0.68
     Islands
    -0.68
     Jav
    -0.68
    bourg
    -0.66
    ãģª
    -0.66
    rities
    -0.66
    POSITIVE LOGITS
     malfunction
    1.49
     ignition
    1.10
     alarm
    1.06
     malf
    1.03
     alarms
    1.02
     shut
    1.00
     activated
    0.94
     shuts
    0.94
    operated
    0.92
     transmitting
    0.91
    Activations Density 0.535%

    No Known Activations