INDEX

Explanations

references to the source or credit of images or other content

oai_token-act-pair · gpt-3.5-turbo

phrases indicating attribution or credit in image descriptions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 withdrawals

-0.67

ashtra

-0.66

 apart

-0.65

 conson

-0.65

 recite

-0.64

 defences

-0.64

 thefts

-0.63

 outper

-0.63

Enabled

-0.63

anooga

-0.63

POSITIVE LOGITS

 STATS

0.86

interstitial

0.79

NBC

0.78

 Wizards

0.77

 LIFE

0.77

 Wikimedia

0.76

zan

0.74

CBS

0.74

 Hogan

0.72

 photographer

0.71

Activations Density 0.045%

references to the source or credit of images or other content

phrases indicating attribution or credit in image descriptions

No Comments

No Known Activations