© 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. 2838
    Prev
    Next
    INDEX
    Explanations

    technical or formal language related to decision-making or division

    oai_token-act-pair · gpt-3.5-turbo

    the word "decided" or its variations and references to decisions or determination

    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
     order
    -0.63
     engine
    -0.59
     engines
    -0.57
    law
    -0.56
     semin
    -0.55
     OM
    -0.55
     Officer
    -0.54
    stones
    -0.54
     woods
    -0.54
     oblig
    -0.53
    POSITIVE LOGITS
    ided
    4.44
    iding
    2.51
    ides
    2.41
    ide
    1.98
    IDES
    1.76
    IDE
    1.67
    iders
    1.64
    isively
    1.62
    ider
    1.54
    isions
    1.43
    Activations Density 0.005%

    No Known Activations