INDEX

Explanations

references to utensils and table setting in dining 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

resave

-0.46

 Keim

-0.44

 Bound

-0.43

Infof

-0.42

 acreage

-0.41

 purchasing

-0.40

 Area

-0.40

 District

-0.39

 area

-0.39

mbggenerated

-0.39

POSITIVE LOGITS

 cutlery

0.88

 spoon

0.84

 spoons

0.76

 Spoon

0.72

spoon

0.70

 cuillère

0.69

 cucchiaio

0.68

Spoon

0.67

プーン

0.66

 cucchia

0.66

Activations Density 0.010%

references to utensils and table setting in dining contexts

No Comments

No Known Activations