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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 0-RES-JB
    11. 4735
    Prev
    Next
    INDEX
    Explanations

    mentions of events, locations, or containers being full or crowded with people or items

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "packed"

    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
    orem
    -0.73
    alf
    -0.72
     attribution
    -0.70
    INGS
    -0.69
    fore
    -0.69
    Mist
    -0.68
    entity
    -0.65
    frog
    -0.65
    aith
    -0.64
     Relations
    -0.63
    POSITIVE LOGITS
     packed
    3.72
    packed
    2.67
     crammed
    2.31
     packing
    1.82
     crowded
    1.68
     jammed
    1.62
     pack
    1.47
     loaded
    1.46
     stacked
    1.46
     filled
    1.43
    Activations Density 0.022%

    No Known Activations