INDEX

Explanations

different types of items, topics, or forms mentioned in a broad context

oai_token-act-pair · gpt-3.5-turbo

references to different categories and types of media, topics, and various forms of content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

MMR

-0.72

uckland

-0.70

pload

-0.68

escription

-0.65

ixties

-0.63

 Ratio

-0.63

xon

-0.62

Ack

-0.62

TW

-0.61

csv

-0.61

POSITIVE LOGITS

 alike

1.11

 respectively

0.93

 thereof

0.92

worldly

0.89

 related

0.79

 besides

0.77

 pertaining

0.77

 indic

0.76

 goodies

0.75

 thereto

0.74

Activations Density 0.215%

different types of items, topics, or forms mentioned in a broad context

references to different categories and types of media, topics, and various forms of content

No Comments

No Known Activations