INDEX

Explanations

phrases or concepts related to numerical figures or estimates

oai_token-act-pair · gpt-3.5-turbo

phrases related to guidelines or rules, particularly in terms of estimation or timing

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

NetMessage

-0.81

JO

-0.76

Attach

-0.74

VPN

-0.72

warts

-0.72

nda

-0.68

JO

-0.67

hij

-0.67

andals

-0.64

andal

-0.64

POSITIVE LOGITS

 averages

1.40

 estimate

1.35

 median

1.28

maximum

1.28

 cutoff

1.28

 average

1.27

 ratio

1.27

 ratios

1.26

 approximate

1.21

 averaged

1.19

Activations Density 1.198%

phrases or concepts related to numerical figures or estimates

phrases related to guidelines or rules, particularly in terms of estimation or timing

No Comments

No Known Activations