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. 0-RES-JB
    6. 17473
    Prev
    Next
    INDEX
    Explanations

    dates or time periods

    oai_token-act-pair · gpt-3.5-turbo

    references to the term "Early" in historical 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
     whatsoever
    -0.67
     racket
    -0.67
     harmless
    -0.67
     orcs
    -0.65
    rael
    -0.65
     spew
    -0.64
     conj
    -0.61
     captured
    -0.61
     Sakuya
    -0.60
    alist
    -0.60
    POSITIVE LOGITS
     Early
    3.33
    Early
    3.06
     Late
    2.15
    early
    1.95
    Late
    1.95
     early
    1.83
     Later
    1.44
     earliest
    1.39
     Earlier
    1.23
     late
    1.20
    Activations Density 0.014%

    No Known Activations