INDEX

Explanations

technical references to coding or programming elements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 betweenstory

-0.78

 ProtoMessage

-0.63

RectangleBorder

-0.63

HomeAsUpEnabled

-0.52

HideFlags

-0.47

олові

-0.46

 BorderSide

-0.46

ViewFeatures

-0.45

BufferException

-0.45

tagext

-0.45

POSITIVE LOGITS

0.52

.")]

0.47

 insurer

0.42

 Ishi

0.42

 asamblea

0.41

 tensors

0.40

 fossil

0.40

 axial

0.39

 institución

0.39

Hosts

0.38

Activations Density 0.006%

technical references to coding or programming elements

No Comments

No Known Activations