Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 370
    Prev
    Next
    INDEX
    Explanations

    phrases beginning with "Whatever"

    oai_token-act-pair · gpt-3.5-turbo

    repeated phrases that signify uncertainty or ambiguity

    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
    aeus
    -0.67
     starting
    -0.66
     twice
    -0.65
     prohibited
    -0.60
     coded
    -0.60
     previously
    -0.60
     only
    -0.60
     replacing
    -0.59
     stopping
    -0.59
     primarily
    -0.58
    POSITIVE LOGITS
    Whatever
    2.75
     Whatever
    2.69
     Regardless
    2.31
    Regardless
    2.27
    Either
    1.97
     Anyway
    1.94
    Anyway
    1.77
     Either
    1.70
    Nevertheless
    1.61
     Nonetheless
    1.57
    Activations Density 0.029%

    No Known Activations