INDEX

Explanations

terms or abbreviations often used for example and illustration

oai_token-act-pair · gpt-3.5-turbo

the use of the term "eg" and related abbreviations that provide examples

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

FINE

-0.74

UNCH

-0.70

Ô

-0.69

Shot

-0.68

SPONSORED

-0.68

 Anthem

-0.67

ãĥĵ

-0.67

UGH

-0.66

ONSORED

-0.66

POSITIVE LOGITS

eg

0.92

ugu

0.85

ocent

0.84

rosso

0.79

omorph

0.76

agons

0.76

orm

0.75

 lapt

0.75

ipt

0.74

uin

0.74

Activations Density 0.007%

terms or abbreviations often used for example and illustration

the use of the term "eg" and related abbreviations that provide examples

No Comments

No Known Activations