INDEX

Explanations

abbreviations or acronyms often associated with official or organizational entities

oai_token-act-pair · gpt-3.5-turbo

abbreviations or acronyms related to organizations and entities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 brill

-0.81

 Incredible

-0.80

 Kardashian

-0.78

ï¸

-0.74

quickShipAvailable

-0.70

 Horus

-0.68

borgh

-0.67

 Sloven

-0.67

ãĥİ

-0.67

 Kanye

-0.66

POSITIVE LOGITS

CS

1.24

CA

1.24

WS

1.20

NP

1.17

SA

1.16

WP

1.15

CP

1.15

GC

1.13

SF

1.11

HS

1.11

Activations Density 0.236%

abbreviations or acronyms often associated with official or organizational entities

abbreviations or acronyms related to organizations and entities

No Comments

No Known Activations