INDEX

Explanations

references to collaborative projects, publications, or activities in various contexts

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

 Fuß

-0.32

CAPIT

-0.31

béco

-0.30

 Photocase

-0.30

 jambes

-0.29

JTable

-0.28

 Sauerstoff

-0.28

 doprav

-0.28

 mukana

-0.28

Erreferentziak

-0.27

POSITIVE LOGITS

 تكبرها

0.70

Personendaten

0.63

 ModelExpression

0.63

IndentedString

0.57

ThroughAttribute

0.54

LLocation

0.54

strijden

0.54

 '\\;'

0.52

Datuak

0.52

 старости

0.51

Activations Density 0.255%

references to collaborative projects, publications, or activities in various contexts

No Comments

No Known Activations