INDEX

Explanations

phrases related to the concept of something being more or less of a certain quality or attribute

oai_token-act-pair · gpt-3.5-turbo

references to increasing quantities or improvements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Travels

-0.78

oes

-0.71

ONES

-0.71

ESE

-0.69

plings

-0.69

ãĥ¡

-0.66

 Nord

-0.66

ovies

-0.65

 Committees

-0.65

 ILCS

-0.64

POSITIVE LOGITS

 complicated

1.01

 than

0.99

 efficient

0.99

than

0.97

 elaborate

0.96

 manageable

0.95

 detailed

0.93

 expensive

0.92

 nuanced

0.92

 stringent

0.91

Activations Density 0.050%

phrases related to the concept of something being more or less of a certain quality or attribute

references to increasing quantities or improvements

No Comments

No Known Activations