© 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. 2971
    Prev
    Next
    INDEX
    Explanations

    comparisons or contrasts between different things

    oai_token-act-pair · gpt-3.5-turbo

    references to various subjects and quantities in the context of comparisons or lists

    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
     Accounting
    -0.71
     Athletics
    -0.70
     Aging
    -0.68
    externalActionCode
    -0.67
    haps
    -0.64
     Gathering
    -0.64
     Buster
    -0.63
    cca
    -0.61
     Tactics
    -0.61
     Norn
    -0.61
    POSITIVE LOGITS
    iest
    1.02
    liest
    0.80
    est
    0.76
    ennes
    0.74
    stuff
    0.72
     alphabet
    0.72
     imaginable
    0.70
     notor
    0.70
    hest
    0.70
    ottest
    0.69
    Activations Density 0.392%

    No Known Activations