INDEX

Explanations

references to NPR (National Public Radio)

oai_token-act-pair · gpt-3.5-turbo

references to news organizations or media outlets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

spring

-0.73

cum

-0.66

ãĤ¹ãĥĪ

-0.65

birth

-0.64

stroke

-0.63

training

-0.63

erate

-0.61

fighter

-0.61

ãĥį

-0.60

jamin

-0.60

POSITIVE LOGITS

NPR

1.54

NPR

1.24

natureconservancy

0.90

 tiss

0.79

 arrang

0.75

DAQ

0.75

£ı

0.73

toc

0.72

 unden

0.71

DEF

0.71

Activations Density 0.005%

references to NPR (National Public Radio)

references to news organizations or media outlets

No Comments

No Known Activations