© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 8-RES-JB
    11. 16875
    Prev
    Next
    INDEX
    Explanations

    instances of the word "Na" followed by a mix of different letters and numbers

    oai_token-act-pair · gpt-3.5-turbo

    references to sodium and its chemical symbol in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    eering
    -0.78
    fare
    -0.76
    tips
    -0.73
     Izan
    -0.71
    rers
    -0.69
    nuts
    -0.69
    ttes
    -0.69
     Journals
    -0.69
     Ange
    -0.66
    papers
    -0.66
    POSITIVE LOGITS
    issance
    1.04
    ïve
    1.04
    ples
    1.01
    uthor
    1.00
    elson
    0.96
    Vi
    0.95
    omi
    0.94
    emonic
    0.93
    vel
    0.92
    emon
    0.92
    Activations Density 0.015%

    No Known Activations