© 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. 8-RES-JB
    11. 3348
    Prev
    Next
    INDEX
    Explanations

    prepositional phrases related to specific actions or methods

    oai_token-act-pair · gpt-3.5-turbo

    various usages of the phrase "the best way to" indicating methods or strategies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Provided
    -0.80
     Appears
    -0.72
    pins
    -0.71
     Klu
    -0.71
     Needs
    -0.69
    lights
    -0.67
    naires
    -0.64
     Conditions
    -0.63
     Balls
    -0.61
    listed
    -0.61
    POSITIVE LOGITS
     communicate
    1.24
     achieve
    1.18
     accomplish
    1.18
     commemorate
    1.16
     differentiate
    1.16
     organize
    1.14
     circumvent
    1.12
     celebrate
    1.12
     visualize
    1.10
     navigate
    1.10
    Activations Density 0.059%

    No Known Activations