INDEX

Explanations

words related to neutrality or being impartial

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 millenn

-0.90

soDeliveryDate

-0.78

Mill

-0.76

teenth

-0.74

Wan

-0.73

hner

-0.70

HAEL

-0.70

Hop

-0.70

omething

-0.70

Amazing

-0.69

POSITIVE LOGITS

izing

1.24

ization

1.16

izers

1.11

izes

1.10

izer

1.08

ized

1.05

ity

1.04

ize

1.00

isation

0.90

ising

0.90

Activations Density 0.017%

words related to neutrality or being impartial

No Comments

No Known Activations