Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 20564
    Prev
    Next
    INDEX
    Explanations

    phrases related to legal procedures and judiciary language

    oai_token-act-pair · gpt-3.5-turbo

    instances of the abbreviation "GE", likely related to General Electric or genetically engineered terms

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Sabb
    -0.73
     rituals
    -0.66
     rub
    -0.64
     abst
    -0.64
     Ob
    -0.62
     pir
    -0.61
     needles
    -0.61
     curly
    -0.60
     tapes
    -0.59
     ul
    -0.59
    POSITIVE LOGITS
    GE
    4.48
    GES
    2.81
     GE
    1.80
    ge
    1.58
    GER
    1.58
    ges
    1.57
    GS
    1.51
    GW
    1.42
    ged
    1.40
    GI
    1.37
    Activations Density 0.011%

    No Known Activations