INDEX

Explanations

first person personal pronouns followed by positive sentiment

oai_token-act-pair · gpt-3.5-turbo

phrases centered around personal experience and perspective

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

batch

-0.82

pak

-0.73

itialized

-0.69

orously

-0.69

vertisements

-0.69

Haunted

-0.69

compatible

-0.68

flush

-0.68

quartered

-0.68

mosp

-0.67

POSITIVE LOGITS

 sake

1.33

 purposes

1.22

selves

0.93

ummies

0.92

 reasons

0.88

 liking

0.82

 personally

0.80

 purpose

0.78

own

0.71

 dear

0.70

Activations Density 0.096%

first person personal pronouns followed by positive sentiment

phrases centered around personal experience and perspective

No Comments

No Known Activations