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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 0-RES-JB
    11. 5162
    Prev
    Next
    INDEX
    Explanations

    special characters, specifically non-English characters

    oai_token-act-pair · gpt-3.5-turbo

    the presence of the character "İ" and terms related to taste or aesthetics, particularly in relation to colors and flavors

    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
     combust
    -0.81
    çĦ
    -0.79
     Bert
    -0.76
     DT
    -0.73
     bust
    -0.72
     bom
    -0.69
     regress
    -0.69
     WARN
    -0.68
     cough
    -0.66
     covenant
    -0.66
    POSITIVE LOGITS
    İ
    2.61
     Pal
    2.15
     palate
    1.58
     palette
    1.58
    Pal
    1.57
    amel
    1.52
    pal
    1.31
     pal
    1.29
     Nou
    1.24
     Sta
    1.13
    Activations Density 0.020%

    No Known Activations