INDEX

Explanations

underscores and uses of wildcard-like patterns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_3/width_16k/average_l0_90

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.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

ſehen

-0.90

parsedMessage

-0.88

 ſind

-0.87

 myſelf

-0.87

 plufieurs

-0.86

 indígen

-0.86

aarrggbb

-0.85

 ainfi

-0.84

 iſt

-0.83

 viſ

-0.83

POSITIVE LOGITS

1.70

(_

1.20

=_

1.17

._

1.17

(_

1.13

/_

1.05

:_

1.02

[_

1.00

//_

1.00

[_

0.95

Activations Density 0.111%

underscores and uses of wildcard-like patterns

No Comments

No Known Activations