INDEX

Explanations

acronyms related to development or authority

oai_token-act-pair · gpt-3.5-turbo

references to Development Authorities and their activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

kies

-0.84

falls

-0.75

ments

-0.71

alter

-0.70

sets

-0.70

ships

-0.70

 Canaver

-0.68

coe

-0.68

naires

-0.67

patrick

-0.67

POSITIVE LOGITS

ILY

1.17

BLE

1.15

VID

1.13

ZZ

0.96

VIS

0.94

WN

0.94

VE

0.87

ZE

0.84

OTA

0.84

ESH

0.83

Activations Density 0.004%

acronyms related to development or authority

references to Development Authorities and their activities

No Comments

No Known Activations