INDEX

Explanations

references to writing skills and activities

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

writing ability

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

ceğini

-0.50

anna

-0.40

migrationBuilder

-0.40

ALLES

-0.40

Skocz

-0.39

 mempel

-0.38

 наи

-0.37

alles

-0.37

culated

-0.36

edil

-0.35

POSITIVE LOGITS

 Writing

0.65

Writing

0.63

 escritura

0.61

 scrittura

0.60

 writing

0.59

écriture

0.54

RenderAtEndOf

0.54

writing

0.51

enderror

0.50

 ModelExpression

0.50

Activations Density 0.006%

references to writing skills and activities

writing ability

No Comments

No Known Activations