INDEX

Explanations

references to changes, particularly in context to policy, documentation, and any specified numerical or categorical classifications

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

 المعيارى

-0.45

 león

-0.41

 desierto

-0.39

 múltiple

-0.38

 cív

-0.38

 ceinture

-0.38

 animación

-0.38

 nativo

-0.37

 abandonado

-0.37

 inconnu

-0.37

POSITIVE LOGITS

UserScript

0.59

 Савезне

0.59

thâu

0.53

 '\\;'

0.52

WriteTagHelper

0.51

 beginnetje

0.51

 Италијани

0.51

NameInMap

0.51

Chham

0.50

хьтан

0.50

Activations Density 0.095%

references to changes, particularly in context to policy, documentation, and any specified numerical or categorical classifications

No Comments

No Known Activations