INDEX

Explanations

numerical references to durations or quantities

oai_token-act-pair · gpt-3.5-turbo

instances of the word "more" indicating additional quantity or continuation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

odor

-0.77

both

-0.73

¬¼

-0.72

=-=-=-=-=-=-=-=-

-0.72

ISTER

-0.70

iman

-0.68

ister

-0.68

anship

-0.66

Gra

-0.65

TextColor

-0.65

POSITIVE LOGITS

 than

1.00

 installments

0.92

 rounds

0.83

 inches

0.79

 layers

0.78

 surprises

0.75

 bedrooms

0.74

 digits

0.73

 weeks

0.73

 months

0.72

Activations Density 0.058%

numerical references to durations or quantities

instances of the word "more" indicating additional quantity or continuation

No Comments

No Known Activations