Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 1-RES-JB
    6. 1538
    Prev
    Next
    INDEX
    Explanations

    locations or features associated with high vantage points

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "overlook" and its variations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    alg
    -0.77
    arter
    -0.75
    agically
    -0.74
    jab
    -0.72
    atom
    -0.69
    ccording
    -0.67
     insepar
    -0.67
     boun
    -0.67
    cel
    -0.66
    omsky
    -0.64
    POSITIVE LOGITS
     overlook
    1.32
     overlooking
    1.18
    ibility
    0.97
    ãĤ¤ãĥĪ
    0.90
    ä¼
    0.82
     misses
    0.81
     overlooked
    0.79
    ibilities
    0.79
    seeing
    0.79
    èĪ
    0.78
    Activations Density 0.014%

    No Known Activations