© 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. 0-RES-JB
    11. 11424
    Prev
    Next
    INDEX
    Explanations

    mentions of different stages during a process or development, such as early stages, final stages, etc

    oai_token-act-pair · gpt-3.5-turbo

    references to various phases or stages of development and progression

    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
    que
    -0.65
    ido
    -0.65
    tu
    -0.63
    uck
    -0.63
    please
    -0.62
    ringe
    -0.62
     Admiral
    -0.61
    iam
    -0.61
     VPN
    -0.60
     Warrant
    -0.60
    POSITIVE LOGITS
     stages
    4.11
     stage
    2.60
     phases
    2.55
    stage
    2.15
     Stage
    2.12
    Stage
    2.09
     steps
    1.43
     phase
    1.38
     waves
    1.32
     arenas
    1.28
    Activations Density 0.012%

    No Known Activations