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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 12518
    Prev
    Next
    INDEX
    Explanations

    phrases or sentences encapsulating a notable quote or remark

    oai_token-act-pair · gpt-3.5-turbo

    punctuation marks, particularly periods and commas, which may indicate the end of statements or dialogue

    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
    Shot
    -0.66
    mit
    -0.65
    kw
    -0.65
    cephal
    -0.64
     prenatal
    -0.64
     Kut
    -0.63
    ient
    -0.63
    eka
    -0.62
    ĻĤ
    -0.62
    holder
    -0.62
    POSITIVE LOGITS
    ushima
    0.70
     ie
    0.69
    izu
    0.68
    VID
    0.61
    £
    0.60
    ãĥĥãĤ¯
    0.59
     Mao
    0.58
    jriwal
    0.58
    ocumented
    0.57
     Whilst
    0.57
    Activations Density 0.052%

    No Known Activations