INDEX

Explanations

formatted elements, particularly related to user interface elements and content specifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.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

expandindo

-0.75

 propOrder

-0.72

 contextLoads

-0.70

 ſhall

-0.69

Ծանոթ

-0.67

SPATH

-0.66

 ſhould

-0.64

 تانيه

-0.64

 Theſe

-0.64

rawDesc

-0.64

POSITIVE LOGITS

clyde

0.52

othea

0.48

そのまま

0.46

himo

0.45

EDEFAULT

0.42

 Unit

0.41

 Daily

0.39

 noDo

0.39

neté

0.38

amerikanische

0.38

Activations Density 0.496%

formatted elements, particularly related to user interface elements and content specifications

No Comments

No Known Activations