© 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. 4-RES-JB
    11. 4256
    Prev
    Next
    INDEX
    Explanations

    terms related to the concept of self, particularly phrases involving self-attributes or self-actions

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to self-destructive behaviors or conditions

    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
    ulhu
    -1.05
     "$:/
    -0.76
     EntityItem
    -0.73
    OUGH
    -0.68
     Springs
    -0.67
     AVG
    -0.67
     Downs
    -0.66
     Massive
    -0.66
     Ashe
    -0.66
    ãĤ¡
    -0.66
    POSITIVE LOGITS
    talk
    1.09
    imposed
    1.07
    generation
    1.03
    conscious
    1.02
    esteem
    0.95
    suff
    0.93
    organ
    0.93
    eating
    0.92
    centered
    0.91
    learning
    0.90
    Activations Density 0.051%

    No Known Activations