INDEX

Explanations

phrases comparing the past to the present or future

oai_token-act-pair · gpt-3.5-turbo

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

ographies

-0.59

eye

-0.58

Pak

-0.56

rich

-0.54

jee

-0.54

chin

-0.53

jiang

-0.53

oris

-0.53

folios

-0.51

letters

-0.51

POSITIVE LOGITS

 earlier

1.18

 yesterday

1.12

 before

1.09

 prior

1.05

 previously

1.04

 beforehand

0.95

 originally

0.92

 during

0.91

 BEFORE

0.87

ufact

0.85

Activations Density 0.184%

phrases comparing the past to the present or future

No Comments

No Known Activations