INDEX

Explanations

phrases related to discussions or comments

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and tracking phrases

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

ROR

-0.77

ELD

-0.74

rolet

-0.70

iph

-0.69

OWN

-0.65

anish

-0.65

ARS

-0.64

ISION

-0.64

¬¼

-0.64

stanbul

-0.64

POSITIVE LOGITS

 chances

1.09

 logically

0.95

 shouldn

0.87

 naturally

0.85

 surely

0.84

 coupled

0.84

 needless

0.81

 understandably

0.79

 excluding

0.78

 please

0.76

Activations Density 0.240%

phrases related to discussions or comments

punctuation marks and tracking phrases

No Comments

No Known Activations