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

    Neuronpedia

    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. 7580
    Prev
    Next
    INDEX
    Explanations

    the name "Bergdahl" with varying degrees of emphasis

    oai_token-act-pair · gpt-3.5-turbo

    references to specific individuals or organizations involved in controversial cases

    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
     Iro
    -0.66
     dup
    -0.65
     dece
    -0.62
     illustrated
    -0.62
     Idlib
    -0.61
     discount
    -0.60
     strips
    -0.60
     tu
    -0.60
     IEEE
    -0.60
     strip
    -0.60
    POSITIVE LOGITS
    erg
    4.82
    ERG
    2.19
    ergy
    1.67
    erk
    1.53
    ergic
    1.51
    erm
    1.34
    erb
    1.27
    urg
    1.27
    err
    1.24
    erman
    1.21
    Activations Density 0.016%

    No Known Activations