Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 687
    Prev
    Next
    INDEX
    Explanations

    the last name "Ross" in various contexts

    oai_token-act-pair · gpt-3.5-turbo

    the term "loss" and its variations in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Interstitial
    -0.76
     originating
    -0.71
     Qian
    -0.64
     NCT
    -0.61
     Hunger
    -0.59
     initiated
    -0.59
     berth
    -0.59
     BET
    -0.59
     White
    -0.58
     cele
    -0.58
    POSITIVE LOGITS
    oss
    4.50
    ossession
    1.99
    OSS
    1.98
    ossus
    1.49
    ose
    1.18
    urn
    1.17
    ross
    1.16
    uss
    1.14
    osa
    1.11
    ack
    1.10
    Activations Density 0.007%

    No Known Activations