© 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 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. 10-RES-JB
    11. 4567
    Prev
    Next
    INDEX
    Explanations

    phrases related to improvement or enhancement

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating an increase in desirability or accessibility

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ovie
    -0.74
    uckle
    -0.73
    minus
    -0.72
    staking
    -0.71
    upuncture
    -0.69
    ENA
    -0.68
    ufact
    -0.66
    liest
    -0.65
    IPS
    -0.63
    outer
    -0.63
    POSITIVE LOGITS
     efficient
    1.23
     attractive
    1.14
     manageable
    1.10
     flexible
    1.07
     comfortable
    1.04
     appealing
    1.03
     aggressive
    1.03
     resilient
    1.02
     sophisticated
    1.00
     equitable
    1.00
    Activations Density 0.108%

    No Known Activations