© 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. 0-RES-JB
    11. 22787
    Prev
    Next
    INDEX
    Explanations

    numerical or alphanumeric sequences, particularly those starting with the letter "F"

    oai_token-act-pair · gpt-3.5-turbo

    the letter "F" in various contexts

    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
    ourke
    -0.65
     vulner
    -0.62
     occas
    -0.62
    aucuses
    -0.62
    isSpecialOrderable
    -0.62
    eatures
    -0.62
    inct
    -0.60
    entimes
    -0.60
    jri
    -0.59
    essage
    -0.58
    POSITIVE LOGITS
     F
    2.99
    F
    2.00
     f
    1.65
    f
    1.42
     FW
    1.32
     FSA
    1.28
     FA
    1.27
     FF
    1.24
     FTA
    1.20
     FS
    1.19
    Activations Density 0.029%

    No Known Activations