© 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. 3-RES-JB
    11. 15547
    Prev
    Next
    INDEX
    Explanations

    the word "hesitate" or its variations in the text

    oai_token-act-pair · gpt-3.5-turbo

    instances of hesitation or indecision

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ham
    -0.84
    model
    -0.76
    olutions
    -0.74
    aled
    -0.73
    rafted
    -0.72
    rollers
    -0.72
    alia
    -0.72
    lining
    -0.71
    andals
    -0.69
    fixed
    -0.69
    POSITIVE LOGITS
     hesitation
    1.23
     hesitate
    1.13
     hesitated
    0.94
     hes
    0.92
     hesitant
    0.79
     shy
    0.78
     wondering
    0.73
     apprehens
    0.72
     warr
    0.72
     apprehension
    0.71
    Activations Density 0.013%

    No Known Activations