INDEX

Explanations

actions related to adding or increasing elements and features

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

phrases related to adding and enhancing features or events

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

adding more or new

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

 régal

-0.58

 perfetta

-0.57

ungkap

-0.57

ngdoc

-0.54

Jeografia

-0.52

stdc

-0.52

 发表于

-0.51

izd

-0.50

BindingResult

-0.50

 préférence

-0.49

POSITIVE LOGITS

 additional

0.90

 hinzu

0.88

 thêm

0.75

 extra

0.74

 layer

0.73

 виправивши

0.72

additional

0.71

 bezeichneter

0.71

AutoresizingMask

0.68

 layers

0.68

Activations Density 0.093%

actions related to adding or increasing elements and features

phrases related to adding and enhancing features or events

adding more or new

No Comments

No Known Activations