INDEX

Explanations

articles and determiners, particularly focusing on the use of 'one' and 'a' in various forms

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_1/width_16k/average_l0_105

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_mlp_out

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

 esterni

-0.72

 kasarigan

-0.68

adpleegd

-0.65

 propOrder

-0.62

tagHelperRunner

-0.60

Hozzáférés

-0.58

 AttributeSet

-0.58

 kaynağından

-0.58

 insuffisamment

-0.57

InjectAttribute

-0.56

POSITIVE LOGITS

1.22

 một

1.10

 μια

0.98

einem

0.96

 eine

0.96

one

0.91

an

0.91

 یک

0.90

einer

0.89

 sebuah

0.88

Activations Density 0.142%

articles and determiners, particularly focusing on the use of 'one' and 'a' in various forms

No Comments

No Known Activations