© 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. 1199
    Prev
    Next
    INDEX
    Explanations

    names of individuals expressing opinions or objections

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "voiced" and its variations

    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
     Schedule
    -0.71
    NRS
    -0.71
    results
    -0.67
    Sche
    -0.65
    onz
    -0.63
     recovery
    -0.63
    mate
    -0.63
    LS
    -0.62
    chel
    -0.61
     permit
    -0.61
    POSITIVE LOGITS
     voiced
    3.77
     voicing
    2.40
     voices
    1.62
     voice
    1.53
     expressed
    1.45
    voice
    1.34
    Voice
    1.29
     articulated
    1.28
     narrated
    1.28
     uttered
    1.27
    Activations Density 0.018%

    No Known Activations