INDEX

Explanations

names, particularly ending with 'os.'

oai_token-act-pair · gpt-3.5-turbo

names and proper nouns

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

 Bentley

-0.97

 Bedford

-0.91

 Mandal

-0.90

 Brend

-0.85

BF

-0.85

 Dudley

-0.85

 Springfield

-0.80

 Carney

-0.80

 Fletcher

-0.79

 Emerson

-0.78

POSITIVE LOGITS

os

2.00

OS

1.70

osa

1.45

osi

1.44

Os

1.39

oses

1.38

os

1.38

ros

1.34

Os

1.33

osaurs

1.33

Activations Density 0.256%

names, particularly ending with 'os.'

names and proper nouns

No Comments

No Known Activations