© 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. 23939
    Prev
    Next
    INDEX
    Explanations

    the word "clearly" followed by either a verb or an adjective

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "clearly" indicating certainty or emphasis

    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
    rella
    -0.85
     lightly
    -0.74
    gently
    -0.73
    eport
    -0.73
    illi
    -0.69
    regate
    -0.68
    tackle
    -0.68
    aily
    -0.68
    Afee
    -0.68
    eries
    -0.67
    POSITIVE LOGITS
     deline
    1.18
     identifiable
    1.11
     marked
    0.95
     defined
    0.93
     distinguish
    0.92
     visible
    0.92
     articulated
    0.89
     differentiated
    0.89
     stated
    0.86
     audible
    0.86
    Activations Density 0.058%

    No Known Activations