INDEX

Explanations

adverbs indicating time such as "now," "currently," and "literally."

oai_token-act-pair · gpt-3.5-turbo

adverbs that indicate frequency or immediacy

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ulence

-0.87

regate

-0.77

 occurs

-0.77

aceutical

-0.69

ortmund

-0.68

 arises

-0.67

 peaks

-0.66

 stems

-0.66

ipeg

-0.65

iox

-0.65

POSITIVE LOGITS

 aware

1.32

 able

1.30

 afraid

1.18

 thankful

1.18

 obsessed

1.16

 interested

1.14

 concerned

1.13

 jealous

1.13

 acquainted

1.10

 grateful

1.10

Activations Density 0.373%

adverbs indicating time such as "now," "currently," and "literally."

adverbs that indicate frequency or immediacy

No Comments

No Known Activations