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

    Neuronpedia

    Jacobian LensNEW
    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. 1398
    Prev
    Next
    INDEX
    Explanations

    short phrases ending with a quote mark

    oai_token-act-pair · gpt-3.5-turbo

    instances of quoted dialogue or statements

    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
    ²¾
    -0.80
    eco
    -0.65
     favourites
    -0.64
    ©¶æ
    -0.64
     scenery
    -0.64
     brill
    -0.63
    ģ«
    -0.62
    ingu
    -0.62
    igun
    -0.61
    erville
    -0.61
    POSITIVE LOGITS
     reads
    1.02
     according
    0.91
     writes
    0.83
    according
    0.78
     says
    0.77
     recalls
    0.76
     said
    0.75
     wrote
    0.74
    said
    0.73
     referring
    0.71
    Activations Density 0.096%

    No Known Activations