INDEX

Explanations

variants of the word "se" and related prefixes and stems, indicating a focus on morphological structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_21/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Theſe

-0.81

 myſelf

-0.78

 itſelf

-0.73

 himſelf

-0.66

ſelf

-0.64

BibitemShut

-0.63

 ſeveral

-0.62

WarningLevel

-0.61

 iſt

-0.60

 purpoſe

-0.58

POSITIVE LOGITS

MigrationBuilder

0.62

bių

0.58

AsUp

0.57

++]=

0.57

 Савезне

0.57

 समीक्षाएं

0.57

ьаж

0.56

 Picchu

0.55

談社

0.55

 يتيمه

0.52

Activations Density 0.095%

variants of the word "se" and related prefixes and stems, indicating a focus on morphological structures

No Comments

No Known Activations