© 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. 1131
    Prev
    Next
    INDEX
    Explanations

    mentions of the name "Ar" with a specific numerical value assigned to it

    oai_token-act-pair · gpt-3.5-turbo

    references to the region or entity identified as "Ar"

    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
     sucked
    -0.78
     lasting
    -0.70
    ictions
    -0.67
    soDeliveryDate
    -0.66
     suck
    -0.66
    wagen
    -0.65
    ussen
    -0.65
     totem
    -0.65
    flix
    -0.62
    emp
    -0.61
    POSITIVE LOGITS
     Ar
    3.36
    Ar
    2.16
     ar
    1.81
     AR
    1.41
     Ard
    1.39
     Ir
    1.37
     Kir
    1.34
     Arn
    1.33
     Arm
    1.32
     Archer
    1.30
    Activations Density 0.020%

    No Known Activations