INDEX

Explanations

high-frequency terms related to significant entities, numbers, or concepts in a text

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

Crain and reporters

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

 Italijani

-0.61

GEBURTSDATUM

-0.58

 pinulongan

-0.56

 resourceCulture

-0.56

ftagPool

-0.55

 Administrativna

-0.54

 betweenstory

-0.53

RenderAtEndOf

-0.52

uxxxx

-0.48

Diweddarwch

-0.47

POSITIVE LOGITS

 reporter

0.52

 <<<<<<<<<<<<<<

0.50

ISupport

0.48

 Reporters

0.48

awtextra

0.47

Rolling

0.47

Voci

0.44

 Rolling

0.44

 Wolken

0.44

 Reporter

0.43

Activations Density 0.001%

high-frequency terms related to significant entities, numbers, or concepts in a text

Crain and reporters

No Comments

No Known Activations