INDEX

Explanations

terms associated with mild health conditions or injuries

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

describing severity or degree

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

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Leder

-0.36

la

-0.32

 joined

-0.30

 trans

-0.29

 partnered

-0.28

 seed

-0.28

Ky

-0.28

 Meier

-0.27

loc

-0.27

CreateModel

-0.27

POSITIVE LOGITS

parsedMessage

0.82

ReusableCell

0.71

ChildScrollView

0.70

 BoxFit

0.65

EDEFAULT

0.63

ofire

0.63

andExpect

0.62

Xna

0.61

enderror

0.61

 InputDecoration

0.60

Activations Density 0.039%

terms associated with mild health conditions or injuries

describing severity or degree

No Comments

No Known Activations