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. 8686
    Prev
    Next
    INDEX
    Explanations

    instances where particular terms are listed together or identified in the text, such as 'or' and 'or', 'or' and 'and', or 'or' and 'letter'

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "or" indicating conditional clauses or alternatives

    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
    ocrats
    -0.82
    ocracy
    -0.78
    istors
    -0.74
    LINE
    -0.71
    ocrat
    -0.70
    Leaks
    -0.67
    orters
    -0.67
    lights
    -0.66
    ires
    -0.66
    ocratic
    -0.66
    POSITIVE LOGITS
    chard
    1.36
    acle
    1.26
    Else
    1.15
    acles
    1.14
    chid
    1.14
    nam
    1.12
     otherwise
    1.10
    ific
    1.10
     alternatively
    1.09
    ifice
    1.09
    Activations Density 0.170%

    No Known Activations