INDEX

Explanations

capital letters followed by periods (e.g., "E.")

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter "e"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mallow

-0.72

 GEAR

-0.70

 Kats

-0.70

 Showdown

-0.68

 Kitchen

-0.68

 Cats

-0.67

 Painter

-0.67

å§«

-0.67

 Noir

-0.66

Kah

-0.65

POSITIVE LOGITS

tymology

1.20

lements

1.18

gypt

1.16

lev

1.15

ighth

1.11

ternally

1.11

ighty

1.08

lder

1.07

astern

1.05

resp

1.04

Activations Density 0.030%

capital letters followed by periods (e.g., "E.")

occurrences of the letter "e"

No Comments

No Known Activations