© 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. 0-RES-JB
    11. 6281
    Prev
    Next
    INDEX
    Explanations

    words related to language or communication, specifically focusing on the use of spoken language

    oai_token-act-pair · gpt-3.5-turbo

    terms related to language and communication, specifically in the context of English-speaking populations

    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
     yard
    -0.72
     temporary
    -0.67
     makeshift
    -0.66
     album
    -0.66
     albums
    -0.65
     fet
    -0.61
     ho
    -0.61
     Stats
    -0.61
     paw
    -0.61
     ramps
    -0.60
    POSITIVE LOGITS
    speaking
    3.46
    important
    1.15
    aye
    1.10
    spe
    1.04
    elected
    0.96
    eric
    0.94
    asking
    0.92
    insured
    0.90
    acus
    0.88
    spoken
    0.87
    Activations Density 0.018%

    No Known Activations