Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 22306
    Prev
    Next
    INDEX
    Explanations

    official websites or platforms that are beginning or starting

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "Begin" and its variations, often indicating the start of events or processes

    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
    lift
    -0.73
    iously
    -0.71
    leton
    -0.66
    amo
    -0.65
    helle
    -0.64
     versa
    -0.62
    Reviewer
    -0.60
    andom
    -0.59
     generator
    -0.58
     RF
    -0.58
    POSITIVE LOGITS
     Begins
    2.73
     Begin
    2.39
    IPS
    1.67
    Begin
    1.48
     Returns
    1.29
     IPS
    1.27
    begin
    1.21
     Covenant
    1.19
     Contin
    1.07
    develop
    1.06
    Activations Density 0.048%

    No Known Activations