INDEX

Explanations

various forms and contexts of the word "work."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_10/width_16k/average_l0_77

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

paravant

-1.06

quelize

-0.95

zbęd

-0.94

 GIPHY

-0.87

CTP

-0.87

 dignité

-0.86

 Enfield

-0.84

 desconhe

-0.84

Datuak

-0.83

 unmute

-0.82

POSITIVE LOGITS

 work

1.83

Work

1.64

 Work

1.64

work

1.62

 works

1.60

 WORK

1.59

WORK

1.51

 Works

1.42

works

1.39

Works

1.35

Activations Density 0.107%

various forms and contexts of the word "work."

No Comments

No Known Activations