© 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
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 4-RES-JB
    6. 10977
    Prev
    Next
    INDEX
    Explanations

    references to proper nouns related to various entities or organizations

    oai_token-act-pair · gpt-3.5-turbo

    references to the letter 'F' followed by numerical digits or other identifiers

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

    No Comments

    Negative Logits
    vier
    -0.73
     Twain
    -0.72
     Aether
    -0.71
     Conquest
    -0.69
     Reborn
    -0.69
     Clever
    -0.68
     Boone
    -0.66
     Heights
    -0.64
     Crusher
    -0.63
     Hades
    -0.63
    POSITIVE LOGITS
    ruits
    1.19
    ocused
    1.17
    ortun
    1.17
    avour
    1.17
    ocusing
    1.16
    ertility
    1.16
    ortunate
    1.12
    andom
    1.09
    isher
    1.09
    ools
    1.08
    Activations Density 0.038%

    No Known Activations