INDEX

Explanations

references to media organizations, particularly NPR

oai_token-act-pair · gpt-3.5-turbo

mentions of NPR and PBS

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

aneously

-0.67

hire

-0.66

gie

-0.64

 Mata

-0.64

wagen

-0.63

 erection

-0.63

 Spit

-0.61

loader

-0.60

clave

-0.60

 Emirates

-0.60

POSITIVE LOGITS

NPR

1.14

NPR

1.08

 Transcript

0.87

 Radio

0.81

 transcripts

0.80

DEF

0.79

 podcasts

0.79

TIT

0.77

OST

0.77

DAQ

0.76

Activations Density 0.013%

references to media organizations, particularly NPR

mentions of NPR and PBS

No Comments

No Known Activations