INDEX

Explanations

terms related to predicates and their usage in logical or mathematical 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

Hour

-0.56

URLException

-0.54

brass

-0.54

 AssemblyVersion

-0.53

Zim

-0.50

Andy

-0.48

Hours

-0.47

SBATCH

-0.47

Months

-0.46

zw

-0.46

POSITIVE LOGITS

 predicate

1.24

 Predicate

1.09

Predicate

1.09

predicate

1.05

 predicates

0.95

predicates

0.79

 predicated

0.63

 TestBed

0.61

 Predic

0.56

 Pradhan

0.54

Activations Density 0.003%

terms related to predicates and their usage in logical or mathematical contexts

No Comments

No Known Activations