INDEX

Explanations

specific names, categories, or important terms related to various subjects

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

Życiorys

-0.68

 disponibilités

-0.66

 فريبيس

-0.62

 AssemblyCulture

-0.60

новништво

-0.54

 MainAxisSize

-0.53

 <<<<<<<<<<<<<<

-0.53

BIÉN

-0.52

Бахар

-0.52

ContentLoaded

-0.51

POSITIVE LOGITS

even

0.40

 doute

0.36

 للمعارف

0.36

 même

0.34

SPJ

0.33

type

0.33

age

0.33

 صوتيه

0.33

 toute

0.33

LAV

0.33

Activations Density 0.297%

specific names, categories, or important terms related to various subjects

No Comments

No Known Activations