INDEX

Explanations

terms indicating a wide scope or range

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 terciopelo

-0.45

 wide

-0.42

 Frankel

-0.39

unggulan

-0.39

mbalikan

-0.38

ppuden

-0.38

 vulnerables

-0.38

 knowing

-0.37

Wide

-0.36

 HUNTER

-0.36

POSITIVE LOGITS

Broad

1.02

 Broad

0.98

 broad

0.97

broad

0.96

 BROAD

0.77

 morning

0.75

 Morning

0.74

Morning

0.74

BROAD

0.70

 MORNING

0.68

Activations Density 0.047%

terms indicating a wide scope or range

No Known Activations