© 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. 4-RES-JB
    11. 2974
    Prev
    Next
    INDEX
    Explanations

    keywords related to the action of reacting or responses

    oai_token-act-pair · gpt-3.5-turbo

    mentions of reactions or responses within a context, particularly related to scientific or urban themes

    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
    holders
    -0.80
    holder
    -0.73
     Treasurer
    -0.61
     Nikon
    -0.61
     Panasonic
    -0.61
    holding
    -0.60
     Starr
    -0.59
     treasurer
    -0.58
     particulars
    -0.58
    draw
    -0.57
    POSITIVE LOGITS
    ency
    0.94
    itect
    0.91
    ivity
    0.87
     Router
    0.83
    halla
    0.82
    ivated
    0.81
    ivation
    0.79
    JS
    0.79
    olver
    0.76
    anooga
    0.75
    Activations Density 0.010%

    No Known Activations