INDEX

Explanations

words that indicate questioning or conditional statements

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

German and other languages

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

cemment

-0.46

fæ

-0.45

tæ

-0.43

 répertoire

-0.43

 oreille

-0.42

Zaw

-0.42

 Jackson

-0.42

 response

-0.41

stø

-0.41

ductory

-0.41

POSITIVE LOGITS

HostException

0.75

Aholisi

0.69

 dass

0.67

makeConstraints

0.66

 bevor

0.60

 να

0.60

Rüyada

0.59

 utafitiHapana

0.59

dass

0.58

abstractmethod

0.57

Activations Density 0.011%

words that indicate questioning or conditional statements

German and other languages

No Comments

No Known Activations