© 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. 16898
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical impact or force, particularly actions involving hitting or striking

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "knocking" and related actions

    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
    ilts
    -0.71
     Direction
    -0.62
     Converted
    -0.61
     modified
    -0.61
    atari
    -0.61
    iked
    -0.60
    held
    -0.60
    ered
    -0.60
    erd
    -0.60
     Income
    -0.59
    POSITIVE LOGITS
     knocking
    3.06
     tearing
    2.09
     banging
    1.97
     slamming
    1.89
     smashing
    1.82
     ripping
    1.79
     pounding
    1.77
     shutting
    1.73
     destroying
    1.68
     crashing
    1.66
    Activations Density 0.043%

    No Known Activations