© 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. 19908
    Prev
    Next
    INDEX
    Explanations

    verbs related to actions performed with a specific intention

    oai_token-act-pair · gpt-3.5-turbo

    terms associated with authority and dominance

    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
    DragonMagazine
    -0.59
     initiative
    -0.54
     impulse
    -0.51
     immersion
    -0.51
     examiner
    -0.51
     inspector
    -0.51
     oldest
    -0.51
     Atlas
    -0.50
     nutshell
    -0.49
     sphere
    -0.49
    POSITIVE LOGITS
    ited
    0.89
    posed
    0.88
    arily
    0.85
    ued
    0.85
    anted
    0.85
    ounced
    0.84
    isted
    0.83
    icked
    0.82
    uated
    0.82
    jured
    0.81
    Activations Density 0.701%

    No Known Activations