INDEX

Explanations

references to top rankings or positions

oai_token-act-pair · gpt-3.5-turbo

references to rankings or positions in a competitive context

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Äĩ

-0.71

 Nare

-0.69

 Consent

-0.68

uld

-0.65

 enqu

-0.64

 Kund

-0.63

Ø©

-0.63

 Correction

-0.63

Ruk

-0.63

microsoft

-0.62

POSITIVE LOGITS

ographical

1.01

 tier

0.97

 scorer

0.91

 contenders

0.90

most

0.88

 rated

0.87

 notch

0.86

ographically

0.86

eka

0.83

 contender

0.81

Activations Density 0.028%

references to top rankings or positions

references to rankings or positions in a competitive context

No Comments

No Known Activations