INDEX

Explanations

references to best practices or standards in various contexts

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

 الرياضيه

-0.69

цездатний

-0.60

fromnode

-0.57

LabelTagHelper

-0.56

MigrationBuilder

-0.55

expandindo

-0.55

ConfigureAwait

-0.52

 يتيمه

-0.52

IntoConstraints

-0.50

ViewFeatures

-0.49

POSITIVE LOGITS

Portale

0.67

AnchorStyles

0.52

Portail

0.51

__(/*!

0.47

 мәкал

0.46

readAs

0.41

 kaynağından

0.40

 Portail

0.39

 Italij

0.38

Tämä

0.38

Activations Density 0.053%

references to best practices or standards in various contexts

No Comments

No Known Activations