INDEX

Explanations

keywords related to news articles, such as names, titles, and positions

oai_token-act-pair · gpt-3.5-turbo

specific names and titles associated with individual persons and their roles

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

 modem

-0.70

 NEXT

-0.66

 defin

-0.65

 cereal

-0.64

 emoji

-0.64

 incub

-0.64

 bothering

-0.64

 record

-0.64

 susp

-0.64

 modifying

-0.63

POSITIVE LOGITS

chan

1.06

sama

1.03

kun

1.01

style

1.01

san

0.95

directed

0.94

based

0.93

cell

0.92

designed

0.91

sen

0.90

Activations Density 0.044%

keywords related to news articles, such as names, titles, and positions

specific names and titles associated with individual persons and their roles

No Comments

No Known Activations