© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 11-RES-JB
    11. 15740
    Prev
    Next
    INDEX
    Explanations

    phrases indicating the start or beginning of an action or feeling

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating the beginning of actions or feelings

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Palest
    -0.80
    leading
    -0.68
    sports
    -0.66
    yet
    -0.66
    boys
    -0.66
    keeping
    -0.64
    until
    -0.64
     continuing
    -0.64
     Marino
    -0.62
     Representatives
    -0.61
    POSITIVE LOGITS
     realise
    1.17
     realize
    1.16
     resemble
    1.12
     accumulate
    1.04
     formulate
    1.04
     worry
    1.01
     unravel
    0.99
     lose
    0.99
     dismantle
    0.99
     rethink
    0.98
    Activations Density 0.056%

    No Known Activations