INDEX

Explanations

significant statements or phrases often related to important findings or conclusions

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

InjectAttribute

-0.57

Vidite

-0.57

 Référence

-0.56

MLLoader

-0.56

 '\\;'

-0.55

stdc

-0.54

delwed

-0.53

 चीज़ों

-0.52

AISSEE

-0.52

featureID

-0.52

POSITIVE LOGITS

 vorstellen

0.37

 chevaux

0.37

 médic

0.37

 attiva

0.35

 deportivo

0.35

 Económica

0.35

 touristique

0.35

 béisbol

0.34

 asesino

0.34

 élas

0.34

Activations Density 0.017%

significant statements or phrases often related to important findings or conclusions

No Comments

No Known Activations