INDEX

Explanations

instances of the word "evaluate" and related terms

oai_token-act-pair · gpt-3.5-turbo

terms related to evaluation and assessment processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Anthem

-0.68

 Flavoring

-0.65

Downloadha

-0.65

Twe

-0.64

href

-0.64

IDA

-0.63

llah

-0.63

İĭ

-0.63

Feel

-0.62

————————————————

-0.62

POSITIVE LOGITS

uating

1.38

uations

1.32

uation

1.27

uated

1.24

uates

1.21

eval

1.21

uate

1.14

uable

1.05

uration

0.88

ined

0.87

Activations Density 0.004%

instances of the word "evaluate" and related terms

terms related to evaluation and assessment processes

No Comments

No Known Activations