INDEX

Explanations

phrases and concepts associated with authority and compliance

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

addContainerGap

-0.80

multer

-0.74

󠁧

-0.67

enumi

-0.63

担

-0.63

 Jefus

-0.62

verläs

-0.62

adone

-0.60

XtraGrid

-0.59

sidemargin

-0.59

POSITIVE LOGITS

 instructions

1.44

 Instructions

1.30

 instructed

1.22

 INSTRUCTIONS

1.20

 instruct

1.18

Instructions

1.17

 instructs

1.17

 directives

1.12

 instruction

1.10

 orders

1.10

Activations Density 0.255%

phrases and concepts associated with authority and compliance

No Comments

No Known Activations