INDEX

Explanations

references to the number of instances and data distribution in datasets

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

 BoxDecoration

-0.48

localctx

-0.48

AnimationsModule

-0.47

contentLoaded

-0.43

///</

-0.43

LookAnd

-0.41

PhysRevLett

-0.41

LEncoder

-0.40

 ostavi

-0.40

writeFieldEnd

-0.40

POSITIVE LOGITS

 ролях

0.40

ülerin

0.38

Built

0.38

 Built

0.37

 resourceCulture

0.36

小姐

0.36

essä

0.35

ayr

0.35

zonder

0.35

 industriel

0.35

Activations Density 0.084%

references to the number of instances and data distribution in datasets

No Comments

No Known Activations