INDEX

Explanations

phrases indicating entitlement or claims to rights or benefits

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

 affari

-0.46

 Moscú

-0.45

 langkah

-0.44

 Arbeit

-0.44

 Baran

-0.43

üyada

-0.43

 Komunikasi

-0.43

 Versammlung

-0.43

Analyst

-0.42

 Schauspielerin

-0.42

POSITIVE LOGITS

entitled

0.92

 entitled

0.90

 entitle

0.87

 entitlement

0.85

 entitlements

0.73

 getTitle

0.63

 setTitle

0.62

title

0.62

Titel

0.60

 Chwiliwch

0.60

Activations Density 0.011%

phrases indicating entitlement or claims to rights or benefits

No Comments

No Known Activations