© 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. 10-RES-JB
    11. 6939
    Prev
    Next
    INDEX
    Explanations

    phrases related to the concept of something happening unexpectedly or miraculously

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "somehow."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    rament
    -0.71
    uties
    -0.71
    emies
    -0.70
    OTOS
    -0.68
    eware
    -0.68
    ourse
    -0.67
    vation
    -0.66
    yip
    -0.65
    emale
    -0.64
    amples
    -0.64
    POSITIVE LOGITS
     somew
    1.49
     managed
    1.15
     magically
    1.10
     mirac
    1.06
     somehow
    1.02
     manages
    1.00
     manage
    0.92
     escaped
    0.91
     survived
    0.88
     forgot
    0.83
    Activations Density 0.028%

    No Known Activations