© 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. 11933
    Prev
    Next
    INDEX
    Explanations

    There is no clear pattern of activation to provide an accurate summary of what neuron 4 is looking for

    oai_token-act-pair · gpt-3.5-turbo

    instances of the character sequence "âĢ" in 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
     torn
    -0.74
     dirt
    -0.70
    anwhile
    -0.70
     Tid
    -0.68
     marching
    -0.68
    lda
    -0.67
     coast
    -0.66
     devastated
    -0.65
     scattering
    -0.65
     nucleus
    -0.65
    POSITIVE LOGITS
    £
    1.08
    ¹
    0.96
    catentry
    0.94
    º
    0.92
    Ķ
    0.91
    ¡
    0.89
    ı
    0.88
    Ĭ
    0.84
    ķ
    0.83
    į
    0.83
    Activations Density 0.758%

    No Known Activations