INDEX

Explanations

references to the "Pyramids."

oai_token-act-pair · gpt-3.5-turbo

references to the word "Py" in various contexts

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

ously

-0.79

DERR

-0.76

nesota

-0.68

hips

-0.68

ILY

-0.66

åŃĲ

-0.65

eous

-0.62

worthiness

-0.62

maid

-0.61

 Hearts

-0.61

POSITIVE LOGITS

ramids

1.42

ramid

1.27

ongyang

1.12

rote

1.00

romancer

0.99

gments

0.93

rene

0.92

rites

0.91

ric

0.91

rex

0.89

Activations Density 0.031%

references to the "Pyramids."

references to the word "Py" in various contexts

No Comments

No Known Activations