INDEX

Explanations

specific terms and phrases related to scientific fields and data structures

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 daze

-0.40

 Höchst

-0.39

 sejauh

-0.39

 buried

-0.38

 Selección

-0.38

 dared

-0.37

 forged

-0.37

 leftover

-0.37

 beheaded

-0.37

 blinded

-0.36

POSITIVE LOGITS

parsedMessage

0.66

DockStyle

0.65

 насељу

0.60

تقاوى

0.60

RectangleBorder

0.59

 становника

0.59

SOUNDBITE

0.59



0.57

EDEFAULT

0.57

CopyWith

0.54

Activations Density 0.052%

specific terms and phrases related to scientific fields and data structures

No Comments

No Known Activations