INDEX

Explanations

technology-related terms or software functionalities

oai_token-act-pair · gpt-3.5-turbo

references to software or application features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

sis

-0.83

oner

-0.75

Ã§a

-0.72

azar

-0.70

cffffcc

-0.70

amn

-0.70

alg

-0.69

apy

-0.69

cel

-0.67

zzi

-0.67

POSITIVE LOGITS

 features

1.05

afety

1.02

Features

1.01

Feature

0.99

 Features

0.97

features

0.92

 prominently

0.83

feature

0.80

 Include

0.77

 enhancements

0.76

Activations Density 0.025%

technology-related terms or software functionalities

references to software or application features

No Comments

No Known Activations