INDEX

Explanations

elements related to typography and font characteristics

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 CanadaChoose

-0.52

艋

-0.51

MigrationBuilder

-0.50

+#+#

-0.49

сылкі

-0.48

ManagedBean

-0.47

parsedMessage

-0.46

dafx

-0.45

 CallOverrides

-0.43

MemoryWarning

-0.43

POSITIVE LOGITS

fonts

0.44

󠁣

0.43

Microkernel

0.42

 nghi

0.40

 pís

0.38

cez

0.37

 schoolboy

0.36

siella

0.35

########.

0.35

 mugs

0.34

Activations Density 0.094%

elements related to typography and font characteristics

No Comments

No Known Activations