INDEX

Explanations

proper nouns, specific names, and terms related to titles or organizations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

 Shyam

-0.78

 Judea

-0.73

ractable

-0.70

GeneratedValue

-0.69

Gom

-0.68

 Rockland

-0.67

httphttps

-0.67

Kast

-0.66

 Henne

-0.66

 Gers

-0.66

POSITIVE LOGITS

Bee

0.88

AnchorStyles

0.87

Loo

0.83

 Neel

0.83

Bee

0.80

Loo

0.80

aloo

0.79

Datuak

0.79

TEE

0.78

noop

0.78

Activations Density 3.144%

proper nouns, specific names, and terms related to titles or organizations

No Comments

No Known Activations