© 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. 7-RES-JB
    11. 3567
    Prev
    Next
    INDEX
    Explanations

    mentions of the name "Charlemagne"

    oai_token-act-pair · gpt-3.5-turbo

    references to a specific character or personality named "Char"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    æĸ¹
    -0.80
     Reloaded
    -0.72
    DERR
    -0.64
    EMS
    -0.64
     AMERICA
    -0.64
    WARD
    -0.64
    CLASSIFIED
    -0.61
     Nex
    -0.60
    ItemTracker
    -0.59
    FactoryReloaded
    -0.58
    POSITIVE LOGITS
    acters
    1.38
    coal
    1.21
    itably
    1.20
    itable
    1.15
    isma
    1.10
    iton
    1.02
    izard
    0.96
    iots
    0.93
    issa
    0.92
    rier
    0.91
    Activations Density 0.030%

    No Known Activations