INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to historical and contemporary organizations or institutions

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

enegger

-0.85

terday

-0.67

abwe

-0.63

theless

-0.61

renheit

-0.60

ONSORED

-0.59

ierrez

-0.58

ÃĥÃĤ

-0.58

milo

-0.58

POSITIVE LOGITS

iest

0.66

 genus

0.64

 axis

0.64

 portion

0.63

osphere

0.62

 matchup

0.59

 universe

0.58

 ecosystem

0.58

 subreddit

0.57

 version

0.56

Activations Density 0.807%

proper nouns

references to historical and contemporary organizations or institutions

No Comments

No Known Activations