© 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. 16462
    Prev
    Next
    INDEX
    Explanations

    words related to magic spells

    oai_token-act-pair · gpt-3.5-turbo
    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
    skirts
    -0.73
     Hudson
    -0.72
     Owens
    -0.71
    payer
    -0.67
     Watts
    -0.67
    TAG
    -0.65
     Roberts
    -0.64
    oros
    -0.64
     Rutherford
    -0.64
    Anderson
    -0.64
    POSITIVE LOGITS
     spell
    3.79
     spells
    2.75
    spell
    2.48
     Spell
    2.13
    Spell
    1.99
     Spells
    1.80
     enchantment
    1.53
     curse
    1.47
     spelling
    1.39
     potion
    1.34
    Activations Density 0.017%

    No Known Activations