© 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 EvalsExportsinterp-engineNEWAPI 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. 12547
    Prev
    Next
    INDEX
    Explanations

    phrases related to continuations or elaborations on a previous statement

    oai_token-act-pair · gpt-3.5-turbo

    clauses that begin with "Still," indicating a shift in perspective or contrast

    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
    minster
    -0.75
    SourceFile
    -0.73
    pecially
    -0.70
    audi
    -0.66
    idia
    -0.65
    ADA
    -0.64
    uran
    -0.63
    oise
    -0.63
    ãĥ³ãĤ¸
    -0.62
    çĶ°
    -0.62
    POSITIVE LOGITS
     there
    0.83
     anecd
    0.81
     suffice
    0.76
     beware
    0.76
     despite
    0.75
     Canaver
    0.74
     nevertheless
    0.70
     nonetheless
    0.68
     according
    0.68
     acknowledging
    0.67
    Activations Density 0.103%

    No Known Activations