INDEX

Explanations

various "Below is" statements with subsequent content

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the existence or presentation of information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 headaches

-0.77

luaj

-0.65

 bells

-0.64

 palms

-0.64

isms

-0.63

 cakes

-0.63

 loves

-0.62

okers

-0.62

eez

-0.62

 sockets

-0.62

POSITIVE LOGITS

 excerpt

0.99

 excerpts

0.83

 transcript

0.82

an

0.80

 summary

0.77

 what

0.73

 footage

0.73

how

0.72

cerpt

0.70

Activations Density 0.084%

various "Below is" statements with subsequent content

phrases indicating the existence or presentation of information

No Comments

No Known Activations