INDEX

Explanations

phrases related to technological items

oai_token-act-pair · gpt-3.5-turbo

occurrences of characters or typographic symbols

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Stats

-0.77

 disadvant

-0.68

 tuna

-0.66

 tradem

-0.63

 mathemat

-0.62

GGGGGGGG

-0.61

 paperback

-0.60

 seiz

-0.60

 gist

-0.60

 Palestin

-0.60

POSITIVE LOGITS

ï¸ı

1.12

spir

1.01

Balt

0.93

0.91

leave

0.87

resh

0.86

Hol

0.85

iversary

0.84

hole

0.84

other

0.83

Activations Density 0.422%

phrases related to technological items

occurrences of characters or typographic symbols

No Comments

No Known Activations