INDEX

Explanations

phrases that denote relationships or connections between concepts

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

хьтан

-0.66

ChildScrollView

-0.65

FMC

-0.54

 noDo

-0.52

 Paglinawan

-0.50

 actionMode

-0.49

FLE

-0.49

 Trom

-0.49

 McGu

-0.48

EmptyString

-0.47

POSITIVE LOGITS

 related

1.13

related

1.13

Related

1.07

 Related

1.02

 RELATED

0.91

 relate

0.89

RELATED

0.86

 relates

0.84

 relation

0.79

relate

0.76

Activations Density 0.028%

phrases that denote relationships or connections between concepts

No Comments

No Known Activations