INDEX

Explanations

terms related to wounds and wound healing

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

wound and healing

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

ⓧ

-0.71

 solito

-0.46

 Autorizaciones

-0.44

 Feier

-0.44

DotNetBar

-0.44

fastjson

-0.43

Pek

-0.43

 Autorisations

-0.43

carpa

-0.42

betrieb

-0.41

POSITIVE LOGITS

 Wound

1.22

Wound

1.20

wound

1.10

 wound

1.08

 Wounded

1.01

 Wounds

0.96

 wounds

0.90

 wounded

0.90

 herida

0.82

 wounding

0.79

Activations Density 0.005%

terms related to wounds and wound healing

wound and healing

No Comments

No Known Activations