© 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. 22584
    Prev
    Next
    INDEX
    Explanations

    references to specific criteria being met or not met

    oai_token-act-pair · gpt-3.5-turbo

    references to standards or benchmarks that are used for evaluation

    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
    joy
    -0.77
    resent
    -0.76
    orld
    -0.76
    ership
    -0.72
    eland
    -0.69
    irth
    -0.69
    ulet
    -0.66
    sky
    -0.64
    standing
    -0.64
    ston
    -0.64
    POSITIVE LOGITS
     criteria
    1.34
     criterion
    1.00
    erion
    0.92
     reviewers
    0.88
    DragonMagazine
    0.83
     guidelines
    0.82
    idelines
    0.77
     thresholds
    0.76
     cutoff
    0.75
    wcsstore
    0.75
    Activations Density 0.012%

    No Known Activations