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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 23122
    Prev
    Next
    INDEX
    Explanations

    the word "completed."

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "completed" and variations of it

    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
    ogy
    -0.79
    udge
    -0.66
    gey
    -0.65
     obscurity
    -0.61
     shaming
    -0.61
    marg
    -0.60
    eter
    -0.60
    culture
    -0.60
    inger
    -0.60
     metaphors
    -0.59
    POSITIVE LOGITS
     completed
    3.52
     finished
    2.23
    Completed
    2.11
     Completed
    2.10
     completion
    2.00
    pleted
    2.00
     completing
    1.95
     completes
    1.94
     complete
    1.78
     finalized
    1.75
    Activations Density 0.008%

    No Known Activations