INDEX

Explanations

references to scandals or controversial events

oai_token-act-pair · gpt-3.5-turbo

references to scandals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

requisite

-0.74

icrobial

-0.72

yip

-0.72

ignt

-0.68

hetics

-0.68

ignty

-0.63

cki

-0.63

emonic

-0.62

erd

-0.61

berman

-0.60

POSITIVE LOGITS

 scandals

1.00

 revolving

0.95

 scandal

0.93

ously

0.92

 involving

0.91

icity

0.84

 plag

0.83

ous

0.80

 allegations

0.79

 engulf

0.76

Activations Density 0.027%

references to scandals or controversial events

references to scandals

No Comments

No Known Activations