© 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. 16688
    Prev
    Next
    INDEX
    Explanations

    words related to image file manipulation techniques

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "gan," indicating a focus on this specific term or related concepts

    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
    İĭ
    -0.75
     differential
    -0.72
     mates
    -0.70
     Topic
    -0.64
    Spanish
    -0.63
     Bravo
    -0.63
    displayText
    -0.62
     Schr
    -0.60
     HMS
    -0.59
    Desc
    -0.58
    POSITIVE LOGITS
    gan
    1.15
    igans
    1.10
    omal
    1.00
    ciating
    0.99
    ogly
    0.96
    za
    0.94
    igan
    0.93
    ecake
    0.92
    non
    0.92
    gha
    0.92
    Activations Density 0.005%

    No Known Activations