INDEX

Explanations

phrases related to show management and user interactions

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

actions and states

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

 препратки

-0.79

boutin

-0.73

󠁴

-0.68

 pulito

-0.66

saraba

-0.66

DockStyle

-0.63

 uitges

-0.63

󠁣

-0.61

 isolato

-0.60

実に

-0.60

POSITIVE LOGITS

 myſelf

0.43

حياته

0.41

 himſelf

0.39

 مؤرشف

0.39

 itſelf

0.38

 élite

0.36

 Lippincott

0.36

 neceſſ

0.35

 respective

0.35

WebAPI

0.35

Activations Density 0.033%

phrases related to show management and user interactions

actions and states

No Comments

No Known Activations