INDEX

Explanations

content related to repetitive actions or phrases

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

efois

-0.57

Diweddarwch

-0.55

 protoimpl

-0.51

setViewportView

-0.51

WebVitals

-0.50

\{\\

-0.50

AndEndTag

-0.49

enumii

-0.49

StoreMessageInfo

-0.48

addCriterion

-0.48

POSITIVE LOGITS

<h2>

0.41

 lenker

0.37

↵↵

0.36

makeText

0.36

!!!

0.36

!!

0.35

AdapterView

0.35

以前

0.35

0.34

buch

0.34

Activations Density 0.020%

content related to repetitive actions or phrases

No Comments

No Known Activations