© 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. 8056
    Prev
    Next
    INDEX
    Explanations

    references to events, individuals, or actions that are being discussed or hypothesized

    oai_token-act-pair · gpt-3.5-turbo

    references to events or actions associated with criminal behavior and investigations

    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
    alg
    -0.58
    antage
    -0.57
     bask
    -0.57
     summarizes
    -0.56
    ggles
    -0.55
     hopefully
    -0.55
    ctors
    -0.53
    cour
    -0.53
     cheers
    -0.53
     Avg
    -0.53
    POSITIVE LOGITS
     improperly
    0.80
     inappropriately
    0.79
     illegally
    0.78
     unlawfully
    0.71
     improper
    0.68
     Kislyak
    0.65
     compromising
    0.64
     paed
    0.63
     nefarious
    0.62
    nikov
    0.62
    Activations Density 1.441%

    No Known Activations