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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 5868
    Prev
    Next
    INDEX
    Explanations

    specific numerical figures and commands related to computer programming

    oai_token-act-pair · gpt-3.5-turbo

    elements related to conditions or requirements for success or participation in a process

    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
    •contextual features
    conditional phrases, often using subjunctive tense. Fires individually on " If"

    No Comments

    Negative Logits
    obbies
    -0.65
     Moroc
    -0.61
    Dub
    -0.59
    ogether
    -0.57
    agra
    -0.56
    jri
    -0.56
     advoc
    -0.56
     ener
    -0.55
     Vaugh
    -0.55
    aeda
    -0.55
    POSITIVE LOGITS
     specified
    0.99
     Cancel
    0.76
     item
    0.74
     optionally
    0.73
     <=
    0.70
     duplicate
    0.68
     parentheses
    0.68
     comma
    0.68
     subparagraph
    0.67
     Item
    0.66
    Activations Density 0.849%

    No Known Activations