INDEX

Explanations

questions and inquiries about understanding and clarification

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

principalTable

-0.90

 ویکی‌پدی

-0.83

ſammen

-0.82

featureID

-0.80

 queſto

-0.80

extAlignment

-0.78

 pinulongan

-0.77

 queſta

-0.75

 ModelExpression

-0.75

BeginContext

-0.73

POSITIVE LOGITS

0.30

0.28

 this

0.28

 Shan

0.27

0.26

0.25

만

0.25

0.24

ilado

0.24

Activations Density 0.079%

questions and inquiries about understanding and clarification

No Comments

No Known Activations