INDEX

Explanations

references to structural analysis and features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 40-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_40/width_16k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.40.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

 SwitchCompat

-0.60

FTAR

-0.45

voirs

-0.44

 oscuros

-0.44

 instruct

-0.43

oughed

-0.43

strument

-0.42

branche

-0.41

strecke

-0.41

струк

-0.41

POSITIVE LOGITS

urally

0.60

#+#

0.51

urals

0.49

+#+

0.45

 betweenstory

0.44

BagLayout

0.43

al

0.41

:✨

0.40

$_['

0.38

 Vikipedi

0.37

Activations Density 0.285%

references to structural analysis and features

No Comments

No Known Activations