INDEX

Explanations

phrases that convey contrast, causation, or consideration

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-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_1/width_16k/average_l0_102

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_resid_post

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

into

-0.49

pu

-0.45

 заслу

-0.44

 became

-0.44

Stop

-0.43

Into

-0.42

 INTO

-0.42

 hold

-0.41

had

-0.41

ADE

-0.40

POSITIVE LOGITS

tagHelperRunner

0.81

 eftersom

0.81

 perquè

0.78

 ponieważ

0.77

 kerana

0.77

 puisque

0.76

 poichè

0.74

 aunque

0.73

 поскольку

0.73

 deoarece

0.72

Activations Density 0.746%

phrases that convey contrast, causation, or consideration

No Comments

No Known Activations