INDEX

Explanations

processes or steps in a sequence

oai_token-act-pair · gpt-3.5-turbo

steps or processes involved in a sequence or procedure

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 lately

-0.89

 underestimated

-0.74

 underest

-0.71

 senses

-0.70

 underestimate

-0.68

violence

-0.67

endon

-0.66

 stupidity

-0.64

BN

-0.64

existent

-0.63

POSITIVE LOGITS

vo

1.02

 whichever

1.00

 prest

0.97

 rinse

0.97

 optionally

0.96

 proceed

0.95

 Repeat

0.94

 restart

0.94

Submit

0.92

 final

0.88

Activations Density 0.795%

processes or steps in a sequence

steps or processes involved in a sequence or procedure

No Comments

No Known Activations