INDEX

Explanations

occurrences of the text "le" in various contexts

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

the occurrences of the French word "le."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_16k/average_l0_119

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.hook_mlp_out

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

le

-2.28

Le

-0.92

les

-0.92

Le

-0.87

le

-0.84

en

-0.72

la

-0.71

-0.70

du

-0.70

te

-0.68

POSITIVE LOGITS

 étoient

1.09

 avoient

1.04

 étoit

1.02

 pouvoit

0.98

 itſelf

0.96

 iſt

0.95

 feroit

0.95

 auroit

0.94

 ſtate

0.88

 ainfi

0.88

Activations Density 0.107%

occurrences of the text "le" in various contexts

the occurrences of the French word "le."

No Comments

No Known Activations