INDEX

Explanations

interrogative phrases and inquiries related to discussions and dialogues

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

 CascadeType

-0.51

 ſtate

-0.48

 myſelf

-0.47

 kaarangay

-0.46

 Administrativna

-0.45

 kasarigan

-0.45

CloseOperation

-0.45

 ſta

-0.43

 faſt

-0.43

 ainfi

-0.43

POSITIVE LOGITS

ftagPool

0.56

SerializeField

0.43

INAS

0.40

ātu

0.39

 atrap

0.39

libft

0.38

 tutur

0.38

úc

0.38

 challenges

0.38

idi

0.38

Activations Density 0.030%

interrogative phrases and inquiries related to discussions and dialogues

No Comments

No Known Activations