INDEX

Explanations

phrases related to creating and discussing content or connections

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 itſelf

-0.45

 faſt

-0.41

 myſelf

-0.41

 Jefus

-0.41

iettivo

-0.40

ſelves

-0.40

 becauſe

-0.40

 ſte

-0.37

 pleaſure

-0.36

 unſ

-0.35

POSITIVE LOGITS

DockStyle

0.71

principalTable

0.62

 يتيمه

0.55

 ویکی‌پدی

0.53

BeginContext

0.50

 AssemblyCulture

0.49

цездатний

0.47

يكب

0.47

الحياه

0.46

 فريبيس

0.46

Activations Density 0.427%

phrases related to creating and discussing content or connections

No Comments

No Known Activations