INDEX

Explanations

commands and obedience-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.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

')],

-0.57

)':

-0.56

 nakalista

-0.54

selves

-0.54

UnsafeEnabled

-0.54

Benjamin

-0.53

quiti

-0.52

 Biôgrafia

-0.52

 cref

-0.52

"]();

-0.52

POSITIVE LOGITS

 obedient

0.82

 obeyed

0.79

 instructions

0.75

 obey

0.69

 blindly

0.67

 obeys

0.67

instructions

0.63

 obeying

0.62

 Instructions

0.62

Instructions

0.62

Activations Density 0.220%

commands and obedience-related terms

No Comments

No Known Activations