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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 19727
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical interactions involving hands

    oai_token-act-pair · gpt-3.5-turbo

    variations of the word "hand" used in different contexts

    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
     disrupted
    -0.60
     Andromeda
    -0.59
    qus
    -0.57
     breached
    -0.57
     peaked
    -0.57
    ynski
    -0.56
     reboot
    -0.56
     starved
    -0.56
     bullied
    -0.56
     Tune
    -0.55
    POSITIVE LOGITS
    hand
    4.43
    hands
    2.37
    Hand
    2.14
    handed
    2.07
    handler
    1.72
    handle
    1.69
     hand
    1.67
    hander
    1.64
     Hand
    1.63
    handled
    1.55
    Activations Density 0.006%

    No Known Activations