INDEX

Explanations

newsletter-related terms and actions

oai_token-act-pair · gpt-3.5-turbo

phrases related to newsletters and their subscription statuses

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

wikipedia

-0.71

bered

-0.63

abama

-0.62

vil

-0.62

 Strauss

-0.62

away

-0.60

amiya

-0.60

governmental

-0.59

 sche

-0.59

bra

-0.58

POSITIVE LOGITS

letters

1.01

letter

0.96

 Subscribe

0.93

 subscribers

0.91

 inbox

0.80

 Newsletter

0.77

LET

0.77

 Emails

0.76

 Sign

0.75

atter

0.74

Activations Density 0.038%

newsletter-related terms and actions

phrases related to newsletters and their subscription statuses

No Comments

No Known Activations