INDEX

Explanations

events or occasions that are the first or initial in a sequence or series

oai_token-act-pair · gpt-3.5-turbo

mentions of events or occurrences related to the word "first."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mbuds

-0.69

hots

-0.65

 somewhere

-0.65

abwe

-0.63

 buddies

-0.63

styles

-0.63

clerosis

-0.63

 depended

-0.63

tics

-0.63

friends

-0.62

POSITIVE LOGITS

 installment

1.13

 edition

0.95

 incarnation

0.93

 chapter

0.89

 iteration

0.88

 portion

0.85

 stages

0.82

 part

0.80

 anniversary

0.80

 phase

0.76

Activations Density 0.255%

events or occasions that are the first or initial in a sequence or series

mentions of events or occurrences related to the word "first."

No Comments

No Known Activations