INDEX

Explanations

statements or arguments related to information or facts

oai_token-act-pair · gpt-3.5-turbo

phrases indicating unusual or noteworthy facts

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

 pione

-0.87

iership

-0.82

ü

-0.78

ā

-0.77

ė

-0.77

÷

-0.77

ù

-0.77

ğ

-0.77

ø

-0.77

POSITIVE LOGITS

 unlike

1.05

 none

1.03

 according

1.00

 neither

1.00

 despite

0.97

 interestingly

0.94

 nobody

0.94

 whereas

0.94

according

0.93

 although

0.91

Activations Density 0.472%

statements or arguments related to information or facts

phrases indicating unusual or noteworthy facts

No Comments

No Known Activations