INDEX

Explanations

entities like names, titles or organizations in a specific format

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and their contexts within citations and quotations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

lling

-0.74

bably

-0.69

arer

-0.67

 sulph

-0.65

 inadequ

-0.63

 forgotten

-0.63

 lifes

-0.63

theless

-0.63

oeuv

-0.63

 concentration

-0.63

POSITIVE LOGITS

BRE

0.93

KT

0.90

FOX

0.89

SCP

0.88

................................................................

0.86

Anonymous

0.86

WF

0.86

Screenshot

0.85

WARNING

0.85

GREEN

0.83

Activations Density 0.131%

entities like names, titles or organizations in a specific format

punctuation marks and their contexts within citations and quotations

No Comments

No Known Activations