INDEX

Explanations

words related to naming conventions or identifiers in a text or code

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "name" and its variants

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

yrinth

-0.92

psey

-0.87

icult

-0.81

ilib

-0.81

elaide

-0.80

romy

-0.78

ĸļ

-0.74

awaru

-0.73

avorite

-0.72

emonium

-0.71

POSITIVE LOGITS

plates

1.28

plate

1.14

paces

1.11

names

0.95

name

0.91

paced

0.86

ames

0.84

pace

0.84

coin

0.80

mark

0.78

Activations Density 0.032%

words related to naming conventions or identifiers in a text or code

occurrences of the word "name" and its variants

No Comments

No Known Activations