INDEX

Explanations

words related to negative connotations

oai_token-act-pair · gpt-3.5-turbo

instances of the prefix "uns" indicating negative or opposite qualities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 nails

-0.64

 Sons

-0.63

 hammer

-0.63

 filing

-0.62

 supers

-0.61

 wills

-0.61

 Rams

-0.60

 coefficient

-0.59

 starters

-0.59

 lions

-0.59

POSITIVE LOGITS

olicited

1.59

avour

1.42

cient

1.38

ustain

1.38

aturated

1.34

chool

1.32

afety

1.28

atisf

1.28

killed

1.26

ight

1.25

Activations Density 0.019%

words related to negative connotations

instances of the prefix "uns" indicating negative or opposite qualities

No Comments

No Known Activations