INDEX

Explanations

instances of reported speech and quotations

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

reported speech with negative context

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

 snippetHide

-0.60

 Meksiku

-0.60

AnchorStyles

-0.55

WireFormatLite

-0.54

oa̍t

-0.54

Jeografia

-0.52

 հղումներ

-0.52

queryInterface

-0.50

Portale

-0.50

abank

-0.49

POSITIVE LOGITS

 warned

0.44

 warning

0.43

 failure

0.40

 état

0.38

 honest

0.38

の原因

0.38

 alarming

0.38

 alarmed

0.38

 blame

0.38

 warn

0.38

Activations Density 0.019%

instances of reported speech and quotations

reported speech with negative context

No Comments

No Known Activations