INDEX

Explanations

references to the concept of "deep" in various contexts, including depth and profoundness

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

+#+#

-0.62

setVerticalGroup

-0.55

 autorytatywna

-0.54

#+#

-0.49

帖最后由

-0.49

expandindo

-0.48

曖昧さ回避

-0.43

########.

-0.43

IUrlHelper

-0.42

تقاوى

-0.42

POSITIVE LOGITS

seated

0.71

rooted

0.68

 seated

0.68

 rooted

0.68

 dive

0.65

 breaths

0.57

 DEEP

0.55

 Dive

0.54

 recesses

0.54

 Deep

0.52

Activations Density 0.085%

references to the concept of "deep" in various contexts, including depth and profoundness

No Comments

No Known Activations