© 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. 591
    Prev
    Next
    INDEX
    Explanations

    time durations expressed as a combination of hours and halves

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "half" or its variations in the text

    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
    imoto
    -0.70
    âĢİ
    -0.70
    ICT
    -0.67
    inus
    -0.66
    piring
    -0.65
    appa
    -0.65
    ct
    -0.65
    create
    -0.64
    ilus
    -0.64
    baugh
    -0.64
    POSITIVE LOGITS
     half
    3.26
    half
    2.60
    Half
    2.41
     Half
    2.25
     halves
    2.22
     quarter
    1.66
     halfway
    1.53
     thirds
    1.41
     HAL
    1.27
    quarter
    1.26
    Activations Density 0.019%

    No Known Activations