INDEX

Explanations

phrases containing contractions such as "Don't" or "It's"

oai_token-act-pair · gpt-3.5-turbo

negative phrases related to self-doubt or inability

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

 acron

-0.69

 Guatem

-0.67

 simultane

-0.66

 decomp

-0.66

 Slav

-0.64

 blanket

-0.64

Osw

-0.63

 SHARES

-0.63

 satell

-0.61

 Alph

-0.60

POSITIVE LOGITS

1.65

tion

1.22

ti

1.21

tan

1.16

tin

1.16

tu

1.14

tions

1.13

tis

1.12

td

1.12

nt

1.06

Activations Density 0.133%

phrases containing contractions such as "Don't" or "It's"

negative phrases related to self-doubt or inability

No Comments

No Known Activations