© 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. 4-RES-JB
    11. 24130
    Prev
    Next
    INDEX
    Explanations

    short phrases ending in 's'

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "is."

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

    No Comments

    Negative Logits
     footh
    -0.62
    CVE
    -0.61
    è¦ļéĨĴ
    -0.60
     scares
    -0.58
     inhibitors
    -0.58
     layoffs
    -0.57
     weights
    -0.57
     bags
    -0.56
     presses
    -0.56
     winters
    -0.56
    POSITIVE LOGITS
    ername
    0.93
    wered
    0.93
    nesday
    0.90
    theless
    0.85
    forth
    0.85
    aurus
    0.77
    pecially
    0.76
    olutely
    0.76
    terday
    0.75
    lightly
    0.73
    Activations Density 0.031%

    No Known Activations