INDEX

Explanations

phrases related to mentions or interactions with the broadcasting network "RT."

oai_token-act-pair · gpt-3.5-turbo

references to radio broadcasts or segments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

recy

-0.70

require

-0.66

flies

-0.65

hire

-0.63

cock

-0.63

gran

-0.63

Ryu

-0.62

iasis

-0.61

hots

-0.60

hiro

-0.59

POSITIVE LOGITS

Ãī

1.30

BF

1.02

LM

0.89

TY

0.89

PC

0.86

Ds

0.84

TN

0.84

irtual

0.83

ICLE

0.82

ivable

0.82

Activations Density 0.045%

phrases related to mentions or interactions with the broadcasting network "RT."

references to radio broadcasts or segments

No Comments

No Known Activations