© 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 EvalsExportsinterp-engineNEWAPI 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. 3-RES-JB
    11. 20809
    Prev
    Next
    INDEX
    Explanations

    references to the country of Greece and its financial situation

    oai_token-act-pair · gpt-3.5-turbo

    references to Greece and its related context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    DERR
    -0.85
    eer
    -0.79
    LESS
    -0.78
    osterone
    -0.78
    \":
    -0.76
    bsp
    -0.76
    alty
    -0.75
    cause
    -0.74
    ////////////////
    -0.73
    ,,,,,,,,
    -0.72
    POSITIVE LOGITS
     Greece
    0.89
     Athens
    0.72
     crisis
    0.71
     bailout
    0.70
     Sicily
    0.70
     Syri
    0.68
     Graveyard
    0.66
    ffiti
    0.66
     devastated
    0.66
    vernment
    0.65
    Activations Density 0.010%

    No Known Activations