INDEX

Explanations

various prefixes and suffixes in words

oai_token-act-pair · gpt-3.5-turbo

instances of profanity and derogatory terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

hyde

-0.93

 Annotations

-0.79

 Macedonia

-0.77

 Rica

-0.76

 Puzzles

-0.74

Ĥİ

-0.73

EStream

-0.71

FactoryReloaded

-0.70

 Fargo

-0.68

å§«

-0.68

POSITIVE LOGITS

iest

1.05

est

1.04

erb

0.89

rep

0.89

ounding

0.88

usive

0.88

uper

0.87

eful

0.87

ib

0.86

ashed

0.85

Activations Density 0.314%

various prefixes and suffixes in words

instances of profanity and derogatory terms

No Comments

No Known Activations