© 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. 17259
    Prev
    Next
    INDEX
    Explanations

    references to legal actions and consequences

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "face" in relation to legal or disciplinary consequences

    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
    ucky
    -0.73
    ary
    -0.71
    aneous
    -0.68
     Paste
    -0.68
    rust
    -0.66
     entit
    -0.65
    rap
    -0.62
    arp
    -0.62
    roma
    -0.61
    rom
    -0.60
    POSITIVE LOGITS
     face
    1.01
     faces
    0.95
    face
    0.86
     Faces
    0.83
    crow
    0.82
     Face
    0.81
    nces
    0.79
    Face
    0.79
    metics
    0.78
    faces
    0.76
    Activations Density 0.023%

    No Known Activations