INDEX

Explanations

proper nouns in the text

oai_token-act-pair · gpt-3.5-turbo

mentions of specific names or figures

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

enza

-0.68

nces

-0.66

cerning

-0.64

heit

-0.62

culated

-0.62

vre

-0.62

phis

-0.62

natureconservancy

-0.62

cially

-0.61

gio

-0.61

POSITIVE LOGITS

icka

0.72

Ïī

0.63

 Ranger

0.63

inski

0.59

Redditor

0.58

Lyn

0.58

EStream

0.58

Row

0.57

Unix

0.56

aroo

0.56

Activations Density 0.348%

proper nouns in the text

mentions of specific names or figures

No Comments

No Known Activations