INDEX

Explanations

attention-grabbing phrases often found in headlines

oai_token-act-pair · gpt-3.5-turbo

content related to headlines from various publications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ensis

-0.74

ajor

-0.73

itaire

-0.73

sterdam

-0.72

ucket

-0.72

gemony

-0.71

rouch

-0.71

heed

-0.70

orum

-0.70

tein

-0.69

POSITIVE LOGITS

 headline

1.00

 headlines

0.95

 proclaiming

0.89

lined

0.79

ãĥŀ

0.68

LINE

0.67

 headlined

0.67

 grabbing

0.67

 proclaim

0.67

vier

0.66

Activations Density 0.065%

attention-grabbing phrases often found in headlines

content related to headlines from various publications

No Comments

No Known Activations