© 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. 3-RES-JB
    11. 10727
    Prev
    Next
    INDEX
    Explanations

    words surrounded by special characters, possibly indicating formatting or structural markers

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of numeric values and formats within the text

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    utics
    -0.64
     undet
    -0.63
     lifes
    -0.60
     wiser
    -0.60
    boro
    -0.59
     hust
    -0.57
     Aval
    -0.54
    utical
    -0.54
    utic
    -0.54
     retreat
    -0.53
    POSITIVE LOGITS
    Meanwhile
    0.84
    Advertisement
    0.83
    SPONSORED
    0.80
    ³³³³
    0.79
    Article
    0.78
    Interestingly
    0.77
    Furthermore
    0.77
    Interest
    0.76
    Ironically
    0.76
    PHOTOS
    0.76
    Activations Density 1.182%

    No Known Activations