Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 8-RES-JB
    6. 3572
    Prev
    Next
    INDEX
    Explanations

    superlatives related to achievements or rankings

    oai_token-act-pair · gpt-3.5-turbo

    numerical rankings and measurements of performance or statistics

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    èĪ
    -0.66
    BAT
    -0.63
     Purs
    -0.60
    Query
    -0.59
    çīĪ
    -0.57
     Penal
    -0.57
    å½
    -0.57
    mar
    -0.56
    Throw
    -0.56
     DeL
    -0.55
    POSITIVE LOGITS
    ever
    0.91
     EVER
    0.88
     anywhere
    0.81
     ever
    0.78
     contiguous
    0.73
     contrace
    0.70
    among
    0.70
    ally
    0.69
    since
    0.69
     shortest
    0.67
    Activations Density 0.211%

    No Known Activations