INDEX

Explanations

instances of specific mathematical or programming terms and relevant contextual phrases

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

 velik

-0.30

ses

-0.30

 reportedly

-0.29

iesc

-0.29

package

-0.29

↵

-0.29

Ses

-0.28

associated

-0.28

 ناب

-0.28

Literatuur

-0.27

POSITIVE LOGITS

httphttps

0.84

 NSCoder

0.76

'{@

0.71



0.68

 صوتيه

0.67

 Normdatei

0.66

 мәкал

0.66

 gynhyrchwyd

0.65

enderror

0.64

 EconPapers

0.63

Activations Density 0.011%

instances of specific mathematical or programming terms and relevant contextual phrases

No Comments

No Known Activations