INDEX

Explanations

phrases indicating loss or transition

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 37-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_37/width_16k/average_l0_124

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.37.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

providedIn

-0.44

BeginContext

-0.38

 PLWABN

-0.33

новниш

-0.33

 endforeach

-0.31

 Mann

-0.31

glVertex

-0.31

matchCondition

-0.30

moveToFirst

-0.29

TargetException

-0.29

POSITIVE LOGITS

 away

3.66

 Away

2.91

Away

2.83

 AWAY

2.69

away

2.69

AWAY

2.33

weg

1.81

aways

1.73

eaway

1.57

taway

1.48

Activations Density 0.437%

phrases indicating loss or transition

No Comments

No Known Activations