INDEX

Explanations

phrases related to performance or functionality

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

as is

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

PerformLayout

-0.57

errHandler

-0.56

+#+

-0.56

 حض

-0.55

-0.53

fjspx

-0.52

 الحره

-0.52

rén

-0.51

Simult

-0.49

 يتيمه

-0.49

POSITIVE LOGITS

As

0.70

as

0.69

as

0.68

As

0.67

Is

0.66

现状

0.65

Is

0.62

is

0.61

IS

0.60

 bestaande

0.59

Activations Density 0.090%

phrases related to performance or functionality

as is

No Comments

No Known Activations