INDEX

Explanations

expressions of apology and regret

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Hauptartikel

-0.79

ließlich

-0.68

GHIJKLM

-0.66

jutant

-0.62

 læng

-0.59

bkz

-0.59

κιν

-0.59

merking

-0.55

ivably

-0.54

ajur

-0.54

POSITIVE LOGITS

 apologies

1.37

 apologise

1.24

 apology

1.21

 apologize

1.18

 sorry

1.13

 apologizing

1.11

 forgive

1.10

 Pardon

1.08

 apologised

1.08

 apologized

1.07

Activations Density 0.078%

expressions of apology and regret

No Comments

No Known Activations