INDEX

Explanations

phrases indicating collaboration or partnerships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_11/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

amentlich

-0.40

saraba

-0.40

 nggak

-0.38

 poř

-0.38

QUENCE

-0.36

 enggak

-0.35

 noqa

-0.34

retudo

-0.34

 очеред

-0.34

 zelve

-0.34

POSITIVE LOGITS

Working

0.91

 Worked

0.84

working

0.84

 trabajó

0.84

 trabaja

0.84

 working

0.82

 Working

0.82

Worked

0.80

 trabajado

0.76

worked

0.75

Activations Density 0.162%

phrases indicating collaboration or partnerships

No Comments

No Known Activations