© 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. 8-RES-JB
    11. 20408
    Prev
    Next
    INDEX
    Explanations

    names of celebrities and characters from popular culture

    oai_token-act-pair · gpt-3.5-turbo

    instances of actor names and notable films or shows

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     vain
    -0.78
     cessation
    -0.74
     gard
    -0.70
     obstruction
    -0.69
     carriage
    -0.68
     appl
    -0.67
    yss
    -0.66
     restored
    -0.65
     manual
    -0.63
     attentive
    -0.63
    POSITIVE LOGITS
    advertising
    1.45
    Probably
    1.07
    Often
    0.97
    Based
    0.96
    Everyone
    0.95
    Sometimes
    0.95
    Most
    0.94
    Honestly
    0.94
    Few
    0.94
    ccording
    0.93
    Activations Density 0.183%

    No Known Activations