INDEX

Explanations

指示信息或提示的词汇和结构。

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

points to following content

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

 للاسماء

-1.02

FunctionFlags

-0.75

StructEnd

-0.75

 betweenstory

-0.74

GraphicsUnit

-0.71

Personendaten

-0.70

 nahilalakip

-0.69

Portály

-0.68

 gynhyrchwyd

-0.66

sizeCache

-0.66

POSITIVE LOGITS

 brief

0.53

 following

0.52

two

0.52

 sample

0.47

inf

0.43

 equation

0.43

 additional

0.42

 three

0.42

 revealing

0.42

0.41

Activations Density 0.026%

指示信息或提示的词汇和结构。

points to following content

No Comments

No Known Activations