INDEX

Explanations

references to publications or documents

oai_token-act-pair · gpt-3.5-turbo

references to published works

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

itialized

-0.83

llan

-0.76

olulu

-0.73

uppet

-0.72

eric

-0.69

heed

-0.69

pell

-0.68

bones

-0.68

othes

-0.68

phas

-0.67

POSITIVE LOGITS

lisher

1.15

lishing

0.88

lishes

0.84

DragonMagazine

0.80

 Publishers

0.74

ä»£

0.73

 Journals

0.72

 publication

0.70

 date

0.70

é»Ĵ

0.70

Activations Density 0.012%

references to publications or documents

references to published works

No Comments

No Known Activations