© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 1-RES-JB
    11. 5090
    Prev
    Next
    INDEX
    Explanations

    the token "kin" followed by a number, with higher activations for "kin10" as opposed to "kin9."

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "kin" and its variations

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Negative Logits
     Seym
    -0.77
    ħĭ
    -0.66
     Debian
    -0.65
     unfocusedRange
    -0.64
     Continued
    -0.63
    ISE
    -0.63
     CoC
    -0.63
     reused
    -0.61
     margin
    -0.61
     Euph
    -0.60
    POSITIVE LOGITS
    kin
    1.18
    etics
    1.11
    ners
    1.07
    etic
    1.06
    etically
    1.01
    osaurs
    0.98
    ning
    0.97
    nings
    0.95
    eman
    0.90
    ny
    0.90
    Activations Density 0.004%

    No Known Activations