© 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. 0-RES-JB
    11. 12451
    Prev
    Next
    INDEX
    Explanations

    realizing or noticing behavior in a sentence

    oai_token-act-pair · gpt-3.5-turbo

    instances of realization or recognition of new insights or truths

    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
    lder
    -0.75
     reside
    -0.73
     serve
    -0.70
    wered
    -0.68
     produce
    -0.67
     send
    -0.66
     qualify
    -0.64
     contribute
    -0.63
    ãĥīãĥ©
    -0.62
     operate
    -0.62
    POSITIVE LOGITS
     realizing
    2.36
     recognizing
    2.26
     noticing
    2.22
     acknowledging
    1.98
     remembering
    1.91
     ignoring
    1.90
     forgetting
    1.88
     discovering
    1.88
     avoiding
    1.86
     rejecting
    1.83
    Activations Density 0.055%

    No Known Activations