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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 0-RES-JB
    6. 5327
    Prev
    Next
    INDEX
    Explanations

    terms related to analysts and analysis

    oai_token-act-pair · gpt-3.5-turbo

    mentions of analysts and their evaluations or opinions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ame
    -0.74
    ress
    -0.71
    boy
    -0.71
    gal
    -0.66
    girl
    -0.65
    ebra
    -0.63
    onto
    -0.63
     knight
    -0.63
     Werewolf
    -0.62
    inian
    -0.62
    POSITIVE LOGITS
     analysts
    3.26
     economists
    2.12
     experts
    1.92
     observers
    1.88
     pundits
    1.86
     commentators
    1.78
     critics
    1.68
     historians
    1.66
     insiders
    1.64
     scholars
    1.62
    Activations Density 0.032%

    No Known Activations