© 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. 14312
    Prev
    Next
    INDEX
    Explanations

    phrases related to composition or makeup

    oai_token-act-pair · gpt-3.5-turbo

    references to components or parts of a whole

    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
    ne
    -0.62
    EE
    -0.59
    ii
    -0.58
    ubb
    -0.57
     valve
    -0.57
     fly
    -0.57
    bye
    -0.57
     pet
    -0.55
     poker
    -0.55
    bury
    -0.55
    POSITIVE LOGITS
     comprises
    1.13
     comprising
    1.10
     comprised
    1.09
    ngth
    0.91
     consists
    0.83
     consisted
    0.82
     comprise
    0.80
    itialized
    0.79
    ivities
    0.79
    clusion
    0.79
    Activations Density 0.007%

    No Known Activations