© 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. 0-RES-JB
    11. 13761
    Prev
    Next
    INDEX
    Explanations

    phrases related to counting or tracking

    oai_token-act-pair · gpt-3.5-turbo

    references to a countdown or numerical sequence

    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
    ornings
    -0.66
     passports
    -0.65
     wa
    -0.65
     tools
    -0.65
    tools
    -0.64
     faux
    -0.62
     undocumented
    -0.61
     passport
    -0.61
     tool
    -0.61
     undisclosed
    -0.61
    POSITIVE LOGITS
    OUNT
    1.43
    SCP
    0.82
    *)
    0.82
    Interstitial
    0.81
    STER
    0.78
    FACE
    0.77
    ISM
    0.76
    KING
    0.75
    Princ
    0.75
    HC
    0.75
    Activations Density 0.000%

    No Known Activations