© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 9-RES-JB
    11. 2016
    Prev
    Next
    INDEX
    Explanations

    the word "shot" or variations of it

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "shot" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 9-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    9
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    herty
    -0.74
    ieties
    -0.70
    éĸ
    -0.69
     streng
    -0.68
    Prin
    -0.66
    urally
    -0.66
    Redditor
    -0.66
    algia
    -0.65
    Case
    -0.63
    rador
    -0.62
    POSITIVE LOGITS
    gun
    1.10
    calling
    1.06
    guns
    0.97
    chens
    0.88
    shot
    0.80
    nikov
    0.80
    call
    0.79
    finder
    0.79
    nery
    0.77
    fire
    0.77
    Activations Density 0.033%

    No Known Activations