© 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. 0-RES-JB
    11. 9757
    Prev
    Next
    INDEX
    Explanations

    mentions of the word "speakers" related to people speaking at events or related to sound devices

    oai_token-act-pair · gpt-3.5-turbo

    references to speakers at events or conferences

    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
    uel
    -0.75
     Deaths
    -0.67
    culosis
    -0.65
     Stard
    -0.65
     Templar
    -0.64
    uild
    -0.64
    ERAL
    -0.63
     Monthly
    -0.63
    aea
    -0.62
    Season
    -0.61
    POSITIVE LOGITS
     speakers
    3.92
     speaker
    3.91
     loudspe
    2.15
     Speaker
    1.88
     microphones
    1.54
     keynote
    1.46
     microphone
    1.45
    spe
    1.45
    Spe
    1.43
     speaking
    1.42
    Activations Density 0.015%

    No Known Activations