© 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. 18446
    Prev
    Next
    INDEX
    Explanations

    the word "checkpoint" and related actions or contexts

    oai_token-act-pair · gpt-3.5-turbo

    references to checkpoints

    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
    soever
    -0.82
    morph
    -0.81
    yss
    -0.73
    tymology
    -0.73
    mort
    -0.73
    mol
    -0.70
    ":"/
    -0.69
    uscript
    -0.68
    WW
    -0.68
    nder
    -0.68
    POSITIVE LOGITS
     checkpoints
    1.30
     checkpoint
    1.15
     clearance
    0.94
     crossings
    0.80
     kios
    0.77
     Patrol
    0.76
     guarding
    0.76
    inery
    0.75
     guards
    0.75
     plaza
    0.74
    Activations Density 0.013%

    No Known Activations