INDEX

Explanations

details or content credits indicated by the word "courtesy"

oai_token-act-pair · gpt-3.5-turbo

references to the word "courtesy."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oshenko

-0.90

akia

-0.88

lear

-0.84

ago

-0.81

olver

-0.80

ako

-0.79

vals

-0.78

atform

-0.78

alone

-0.78

agh

-0.77

POSITIVE LOGITS

tesy

1.02

 courtesy

0.94

 swipe

0.76

 Thomson

0.73

rored

0.72

 Credit

0.70

 Cullen

0.69

giving

0.68

CLASSIFIED

0.67

 PHOTO

0.66

Activations Density 0.006%

details or content credits indicated by the word "courtesy"

references to the word "courtesy."

No Comments

No Known Activations