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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 0-RES-JB
    6. 23627
    Prev
    Next
    INDEX
    Explanations

    technical terms or instructions that involve extensions or modifications

    oai_token-act-pair · gpt-3.5-turbo

    references to specific entities or terminologies related to external functions or conditions

    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
     helm
    -0.74
     CNBC
    -0.71
     muster
    -0.71
     Ballard
    -0.69
    ny
    -0.69
     lottery
    -0.69
     fle
    -0.68
     scrimmage
    -0.66
     luck
    -0.65
    sted
    -0.65
    POSITIVE LOGITS
    Ext
    3.14
     Extensions
    1.32
    Reb
    1.30
    Extra
    1.26
    Opt
    1.25
    Assembly
    1.17
    Expl
    1.14
    Cos
    1.13
    Root
    1.12
    Alternative
    1.10
    Activations Density 0.034%

    No Known Activations