INDEX

Explanations

references to authority figures or titles, particularly in formal or medical contexts

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

 ModelExpression

-0.47

bray

-0.44

 lenker

-0.42

awtextra

-0.42

Board

-0.42

BagLayout

-0.40

enschappelijke

-0.40

 Board

-0.40

 NSCoder

-0.40

findpost

-0.40

POSITIVE LOGITS

bu

2.17

Bu

2.08

Bu

2.00

bu

1.85

 Бу

1.76

Buc

1.76

BU

1.74

 бу

1.70

Buk

1.66

BU

1.66

Activations Density 1.274%

references to authority figures or titles, particularly in formal or medical contexts

No Comments

No Known Activations