INDEX

Explanations

instances of the word "Related" and references to association or connections

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

arer

-0.80

acers

-0.75

arest

-0.74

UGE

-0.73

amping

-0.68

»Ĵ

-0.67

elfth

-0.66

ohm

-0.66

agall

-0.65

ample

-0.65

POSITIVE LOGITS

 Stories

0.90

 Links

0.85

 Topics

0.84

Related

0.81

 Articles

0.80

 Laugh

0.80

Sym

0.78

 Factors

0.76

Images

0.76

 Cosponsors

0.75

Activations Density 0.020%

related articles or content sections

instances of the word "Related" and references to association or connections

No Comments

No Known Activations