INDEX

Explanations

Keywords related to news or journalism, such as "story" or "reporting"

oai_token-act-pair · gpt-3.5-turbo

references to stories

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

ignt

-0.83

aez

-0.82

amily

-0.81

ategory

-0.81

achev

-0.77

orneys

-0.77

heed

-0.76

alus

-0.76

inence

-0.75

ensable

-0.75

POSITIVE LOGITS

telling

1.18

te

0.94

arc

0.91

board

0.91

boards

0.90

book

0.86

 arcs

0.85

 revolving

0.82

books

0.81

 reprinted

0.80

Activations Density 0.052%

Keywords related to news or journalism, such as "story" or "reporting"

references to stories

No Comments

No Known Activations