INDEX

Explanations

phrases that convey consistent patterns or behaviors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 للمعارف

-0.62

حياته

-0.58

parsedMessage

-0.56

 informée

-0.56

LLocation

-0.54

awtextra

-0.54

 vooz

-0.54

RenderAtEndOf

-0.53

 メンテナ

-0.53

حياتها

-0.53

POSITIVE LOGITS

 always

1.05

always

0.98

 ALWAYS

0.94

Always

0.93

 Always

0.93

 siempre

0.93

 zawsze

0.87

 תמיד

0.85

 sempre

0.84

 alltid

0.84

Activations Density 0.379%

phrases that convey consistent patterns or behaviors

No Comments

No Known Activations