INDEX

Explanations

phrases indicating specificity or particularity in context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

蚪

-0.54

omatic

-0.45

存于互联网档案馆

-0.45

 múltiple

-0.44

multirow

-0.42

Station

-0.42

揄

-0.41

Klar

-0.41

titutes

-0.41

 alluminio

-0.41

POSITIVE LOGITS

 particularly

1.13

 especially

1.06

 Particularly

0.98

particularly

0.98

Especially

0.95

especially

0.93

 szczególnie

0.90

 especialmente

0.90

 PARTICULAR

0.90

 Especially

0.89

Activations Density 0.275%

phrases indicating specificity or particularity in context

No Comments

No Known Activations