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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 3-RES-JB
    11. 20799
    Prev
    Next
    INDEX
    Explanations

    references to weather phenomena, particularly fog

    oai_token-act-pair · gpt-3.5-turbo

    references to fog and related obscuring elements

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    brate
    -0.72
    realDonaldTrump
    -0.67
    Murray
    -0.66
     Parenthood
    -0.66
    Comments
    -0.65
     VIDEOS
    -0.65
    ensional
    -0.65
    brates
    -0.65
    Union
    -0.65
    IGN
    -0.64
    POSITIVE LOGITS
     fog
    1.16
    ging
    1.05
    gers
    1.01
    gy
    0.98
    glers
    0.98
    ged
    0.93
     cloud
    0.90
    rain
    0.87
     clouds
    0.87
     haze
    0.83
    Activations Density 0.006%

    No Known Activations