INDEX

Explanations

statements that define or describe concepts and entities

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

definitions and descriptions

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

 estekak

-0.69

expandindo

-0.66

Geplaatst

-0.66

 bezeichneter

-0.62

 ब्रेकडाउन

-0.61

URLException

-0.60

uxxxx

-0.58

ंदीखरीदारी

-0.58

évaluateur

-0.57

Jereo

-0.56

POSITIVE LOGITS

 charité

0.36

Backdrop

0.35

darah

0.34

îné

0.33

 hierba

0.33

 family

0.32

 vibra

0.32

 Batalla

0.32

 whoſe

0.31

 intende

0.31

Activations Density 0.022%

statements that define or describe concepts and entities

definitions and descriptions

No Comments

No Known Activations