© 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. 4-RES-JB
    11. 6585
    Prev
    Next
    INDEX
    Explanations

    references to social or formal events

    oai_token-act-pair · gpt-3.5-turbo

    mentions of "reception" in various contexts

    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
    chn
    -0.75
     mortg
    -0.72
    ffe
    -0.62
     circling
    -0.62
    holder
    -0.61
     diabetic
    -0.61
    userc
    -0.60
    hed
    -0.60
    hover
    -0.59
     Sul
    -0.58
    POSITIVE LOGITS
     reception
    1.09
    aries
    0.95
    ally
    0.93
    alled
    0.90
    DragonMagazine
    0.87
    naire
    0.84
    anche
    0.81
    arity
    0.81
    itory
    0.80
     Kirin
    0.79
    Activations Density 0.023%

    No Known Activations