© 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. 1-RES-JB
    11. 23206
    Prev
    Next
    INDEX
    Explanations

    historical events or initiatives that have been officially authorized or funded by a higher authority

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "commissioned" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ox
    -0.75
    Sche
    -0.75
    corn
    -0.73
    kun
    -0.71
    xa
    -0.70
    amac
    -0.70
    Cube
    -0.68
    va
    -0.67
    um
    -0.67
    Bah
    -0.67
    POSITIVE LOGITS
     commissioned
    1.46
     commission
    0.96
     commissions
    0.94
    rontal
    0.84
     destro
    0.80
     confir
    0.78
    structed
    0.77
     nudity
    0.75
     councill
    0.74
     Bride
    0.74
    Activations Density 0.005%

    No Known Activations