INDEX

Explanations

numerical data and associated units or percentages

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

-0.58

dorf

-0.58

 pied

-0.57

 poch

-0.57

kut

-0.56

 inne

-0.56

 sotto

-0.56

otten

-0.54

POSITIVE LOGITS

SBATCH

0.71

ైన

0.68

 vidare

0.67

ื่อน

0.64

 perry

0.63

kuuta

0.63

 član

0.62

 religieuses

0.62

SdkVersion

0.62

텝

0.61

Activations Density 0.296%

numerical data and associated units or percentages

No Comments

No Known Activations