INDEX

Explanations

specific scientific terms and symbols related to experimental conditions and measurement outcomes

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

parenthesized abbreviations

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

expandindo

-0.68

 propOrder

-0.58

featureID

-0.52

 AspNetCore

-0.50

Autoritní

-0.50

جوايز

-0.49

enderror

-0.48

casian

-0.48

ंदीखरीदारी

-0.47

 المعيارى

-0.47

POSITIVE LOGITS

velles

0.40

 hence

0.39

Rez

0.35

záll

0.35

ensiva

0.35

 nią

0.34

 Bewußt

0.34

 named

0.33

BeginContext

0.33

şekkür

0.32

Activations Density 0.237%

specific scientific terms and symbols related to experimental conditions and measurement outcomes

parenthesized abbreviations

No Comments

No Known Activations