© 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. 2-RES-JB
    11. 9102
    Prev
    Next
    INDEX
    Explanations

    words containing the subsequence "gan" with a strong preference for "egan"

    oai_token-act-pair · gpt-3.5-turbo

    mentions of the word "gan" in various forms and contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     differential
    -0.75
     Topic
    -0.73
    İĭ
    -0.67
     Bravo
    -0.60
     contagious
    -0.60
     Jordanian
    -0.60
     mates
    -0.59
    Spanish
    -0.59
     pessimistic
    -0.58
    PER
    -0.58
    POSITIVE LOGITS
    gan
    1.42
    igans
    1.15
    igan
    1.02
    ogly
    0.99
    ciating
    0.95
    emon
    0.91
    isbury
    0.90
    wen
    0.90
    culosis
    0.90
    ghan
    0.88
    Activations Density 0.009%

    No Known Activations