INDEX

Explanations

keywords and phrases related to evidence and testimony

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

long time

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

Geplaatst

-0.64

Hochspringen

-0.62

IsContent

-0.57

énario

-0.57

AndEndTag

-0.56

 protoimpl

-0.53

.*")]

-0.53

 sumpay

-0.52

complexContent

-0.52

MessageTagHelper

-0.52

POSITIVE LOGITS

 zabud

0.32

anInt

0.32

wicklung

0.31

fb

0.31

 Administrativna

0.30

 goederen

0.30

хьтан

0.30

utilisons

0.30

ніципа

0.30

舺

0.30

Activations Density 0.013%

keywords and phrases related to evidence and testimony

long time

No Comments

No Known Activations