© 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. 2-RES-JB
    11. 4492
    Prev
    Next
    INDEX
    Explanations

    mentions of the word "electrocution."

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "elect"

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Lens
    -0.76
    BuyableInstoreAndOnline
    -0.72
    ãĤ¬
    -0.70
    acca
    -0.69
    AMI
    -0.69
    MpServer
    -0.68
    Princ
    -0.67
    TAG
    -0.67
     Flavoring
    -0.66
     HMS
    -0.63
    POSITIVE LOGITS
    eering
    1.04
    elect
    0.95
     elect
    0.92
    roc
    0.91
    rification
    0.88
    icter
    0.86
    inations
    0.86
    rated
    0.85
    dinand
    0.84
    vind
    0.84
    Activations Density 0.007%

    No Known Activations