INDEX

Explanations

categories related to buildings and architecture

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_18/width_65k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.18.hook_mlp_out

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

 EconPapers

-0.83

XmlAccessorType

-0.71

بوابة

-0.67

adaptiveStyles

-0.63

Portail

-0.60

exitRule

-0.58

medriver

-0.56

 isNameExpr

-0.56

hyrchwyd

-0.56

saraba

-0.55

POSITIVE LOGITS

 Real

0.55

 real

0.52

omene

0.49

+:+

0.49

Ull

0.48

Real

0.48

 Isra

0.47

real

0.46

Realtime

0.45

Pr

0.45

Activations Density 0.154%

categories related to buildings and architecture

No Comments

No Known Activations