INDEX

Explanations

comparisons and descriptions of similarity and difference among elements or concepts

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

sameness or similarity

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

protetor

-0.34

 Grüsse

-0.31

 frü

-0.30

 itself

-0.30

setFill

-0.29

 beauty

-0.29

accompag

-0.29

HDC

-0.28

ogue

-0.28

 Stock

-0.28

POSITIVE LOGITS

featureID

0.88

 ویکی‌پدی

0.81

MigrationBuilder

0.73

ConstraintMaker

0.69

kháu

0.69



0.67

GTCX

0.67

 noDo

0.65

 мәкал

0.65

IntoConstraints

0.65

Activations Density 0.068%

comparisons and descriptions of similarity and difference among elements or concepts

sameness or similarity

No Comments

No Known Activations