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 · Initial Gemma-2B-IT SAEs
    3. Gemma-2B-IT
    4. Residual Stream
    5. 12-RES-JB
    6. 10009
    Prev
    Next
    INDEX
    Explanations

    text related to detailed descriptions of items or actions

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot

    contextual cues or transitions in text, particularly those indicating the start of explanations, definitions, or important points in a passage.

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @eekay

    text that describes details or characteristics, particularly in the form of descriptive sentences or paragraphs.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @eekay
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/Gemma-2b-IT-Residual-Stream-SAEs/gemma_2b_it_blocks.12.hook_resid_post_16384
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    HuggingFaceFW/fineweb
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.12.hook_resid_post
    Hook Layer
    12
    Architecture
    standard
    Context Size
    1,024
    Dataset
    Skylion007/openwebtext
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    1150
    +0.08
    0.2%
    1445
    +0.08
    0.2%
    1056
    +0.08
    0.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    1959
    +0.08
    0.12
    1885
    +0.08
    0.04
    1380
    +0.08
    0.00
    Negative Logits
     embra
    -2.29
     effe
    -2.26
     overla
    -2.24
     oner
    -2.23
     dispen
    -2.23
     igno
    -2.16
     abnorm
    -2.16
     nutr
    -2.15
     pessi
    -2.12
     inder
    -2.12
    POSITIVE LOGITS
    <bos>
    1.09
     due
    0.78
    contentLoaded
    0.72
    ..
    0.71
    <eos>
    0.70
    ;
    0.69
    )
    0.67
    。
    0.67
    .
    0.66
    和
    0.66
    Activations Density 2.364%

    No Known Activations