© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 9543
    Prev
    Next
    INDEX
    Explanations

    phrases related to making changes or modifications

    oai_token-act-pair · gpt-3.5-turbo

    the infinitive form "to" followed by verbs indicating actions or changes

    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
    croft
    -0.72
    arten
    -0.71
    furt
    -0.68
    lance
    -0.64
    need
    -0.62
    soDeliveryDate
    -0.62
    sle
    -0.61
    eele
    -0.61
    é¾įåĸļ士
    -0.60
    igmatic
    -0.60
    POSITIVE LOGITS
     ensure
    0.96
     improve
    0.90
     maximize
    0.89
     compensate
    0.85
     overcome
    0.85
    iling
    0.85
    asted
    0.84
    wered
    0.81
    tnc
    0.80
     reduce
    0.80
    Activations Density 0.241%

    No Known Activations