INDEX

Explanations

references to the positioning or dimensions of elements, specifically related to the "bottom" aspect

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_4/width_65k/average_l0_66

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.4.hook_mlp_out

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bottom

-1.49

Bottom

-1.48

 Bottom

-1.38

 bottom

-1.33

 BOTTOM

-1.09

BOTTOM

-1.08

 bottoms

-0.84

botto

-0.81

Top

-0.79

top

-0.77

POSITIVE LOGITS

tagHelperRunner

0.73

"]];

0.67

"]);

0.66

TagMode

0.65

 تانيه

0.65

adaptiveStyles

0.63

Enllaces

0.63

 estekak

0.60

"]);

0.59

Autoritní

0.59

Activations Density 0.012%

references to the positioning or dimensions of elements, specifically related to the "bottom" aspect

No Comments

No Known Activations