INDEX

Explanations

references to academic journals and their articles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.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

 możliwe

-0.62

setAll

-0.58

...

-0.54

ėl

-0.54

✪

-0.52

<eos>

-0.52

 prohibido

-0.52

 måte

-0.51

Shape

-0.50

 Mustafa

-0.50

POSITIVE LOGITS

 Jamb

1.15

 jLabel

1.07

 Jays

1.07

Jop

1.04

Jou

1.03

Jy

1.03

JAS

1.02

Jays

1.00

Jav

1.00

jy

0.99

Activations Density 0.663%

references to academic journals and their articles

No Comments

No Known Activations