INDEX

Explanations

numbers mentioned in text

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "numbers."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

nee

-0.74

Interstitial

-0.72

ModLoader

-0.72

 Shroud

-0.70

POSE

-0.68

INTON

-0.68

clair

-0.66

struction

-0.66

Cast

-0.65

ISSION

-0.64

POSITIVE LOGITS

 numbers

1.03

 Numbers

0.94

paces

0.85

 digits

0.82

pace

0.81

omial

0.80

peak

0.80

peed

0.78

 percentages

0.76

 averages

0.75

Activations Density 0.019%

numbers mentioned in text

repeated mentions of the word "numbers."

No Comments

No Known Activations