INDEX

Explanations

specific terms related to medical or physiological concepts and their properties

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

approach, use, because

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

parsedMessage

-0.77

webElement

-0.71

sizeCache

-0.70

DoubleQuotes

-0.70

IntoConstraints

-0.69

AddTagHelper

-0.69

 ब्रेकडाउन

-0.68

setVerticalGroup

-0.68

featureID

-0.66

IndentedString

-0.66

POSITIVE LOGITS

 Grunde

0.38

ավ

0.32

use

0.32

 because

0.31

 olma

0.30

 ام

0.30

 réserver

0.29

 permettre

0.29

ายุ

0.28

 tué

0.28

Activations Density 0.307%

specific terms related to medical or physiological concepts and their properties

approach, use, because

No Comments

No Known Activations