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. 4-RES-JB
    6. 15107
    Prev
    Next
    INDEX
    Explanations

    words or phrases related to being physically stuck or confined

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "trapped" in various contexts, especially relating to confinement or danger

    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
    issance
    -0.83
    esan
    -0.79
    sburgh
    -0.78
    edly
    -0.74
    eds
    -0.73
    sych
    -0.70
    edy
    -0.67
    sburg
    -0.66
    itage
    -0.66
    ed
    -0.64
    POSITIVE LOGITS
     trapped
    0.85
    ©¶æ¥µ
    0.77
     Tracks
    0.76
    untarily
    0.74
     trapping
    0.74
     squid
    0.71
    ngth
    0.67
     souls
    0.67
    oxin
    0.66
    iven
    0.66
    Activations Density 0.019%

    No Known Activations