INDEX

Explanations

references to related content and suggestions for further reading

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

概要

-0.54

 OFDb

-0.47

beforeEach

-0.45

-0.44

copyTo

-0.42

and

-0.40

-0.39

 chiaramente

-0.39

 wherever

-0.39

POSITIVE LOGITS

 Related

1.24

Related

1.21

 related

1.20

related

1.06

 read

1.05

Also

0.98

 also

0.97

RELATED

0.97

 READ

0.95

 Also

0.94

Activations Density 0.087%

references to related content and suggestions for further reading

No Comments

No Known Activations