INDEX

Explanations

introductory phrases before a specific action or explanation

oai_token-act-pair · gpt-3.5-turbo

punctuation and commas in writing

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

arch

-0.69

minster

-0.68

bryce

-0.63

etheless

-0.60

rider

-0.60

 brill

-0.60

ogly

-0.59

arov

-0.56

utor

-0.56

<@

-0.55

POSITIVE LOGITS

 please

0.91

 however

0.88

please

0.74

iegel

0.73

 multiply

0.68

 Flavoring

0.65

 CLICK

0.64

 preferably

0.64

 Kessler

0.63

 Collider

0.63

Activations Density 0.470%

introductory phrases before a specific action or explanation

punctuation and commas in writing

No Comments

No Known Activations