INDEX

Explanations

character sequences that are often associated with high intensity or impact

oai_token-act-pair · gpt-3.5-turbo

verbs related to actions and events that carry significant implications or effects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 dressing

-0.69

 praying

-0.68

 listening

-0.67

 pretending

-0.65

 complying

-0.65

 attending

-0.65

 apologizing

-0.63

 withdrawing

-0.62

 rebellious

-0.61

 behaving

-0.61

POSITIVE LOGITS

ifies

1.15

uates

1.12

izes

1.07

iates

0.93

itates

0.91

shed

0.90

ulates

0.88

ersed

0.88

tains

0.87

ceed

0.86

Activations Density 0.195%

character sequences that are often associated with high intensity or impact

verbs related to actions and events that carry significant implications or effects

No Comments

No Known Activations