INDEX

Explanations

words related to associations or connections between concepts

oai_token-act-pair · gpt-3.5-turbo

phrases indicating associations or relationships

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

emouth

-0.66

 priced

-0.64

abases

-0.63

schild

-0.63

Reviewed

-0.60

published

-0.60

 premiered

-0.59

STAT

-0.59

ucci

-0.57

 apologized

-0.57

POSITIVE LOGITS

 regards

1.00

 regard

0.79

è¦ļéĨĴ

0.75

ä½ľ

0.72

 extinction

0.70

emonic

0.69

stood

0.65

ryan

0.61

åİ

0.61

ript

0.61

Activations Density 0.069%

words related to associations or connections between concepts

phrases indicating associations or relationships

No Comments

No Known Activations