INDEX

Explanations

references to structured data or programming constructs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_11/width_16k/average_l0_80

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 usah

-0.54

 onOptions

-0.54

 tearing

-0.53

はじめに

-0.51

oneofs

-0.51

 isInitialized

-0.50

trex

-0.50

 Italijani

-0.49

лети

-0.49

cessite

-0.49

POSITIVE LOGITS

principalColumn

0.73

multer

0.65

extAlignment

0.60

enumi

0.58

enumii

0.58

AxisAlignment

0.56

 оригіналу

0.56

 ويكيميديا

0.55

IsContent

0.53

انيف

0.52

Activations Density 0.438%

references to structured data or programming constructs

No Comments

No Known Activations