INDEX

Explanations

phrases related to doors and transition points

oai_token-act-pair · gpt-3.5-turbo

phrases related to opportunities and access

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

Ð»

-0.74

oday

-0.73

oice

-0.71

onna

-0.70

 à¤

-0.70

 WITHOUT

-0.67

hetical

-0.66

ually

-0.65

requent

-0.65

wcsstore

-0.65

POSITIVE LOGITS

agall

0.77

 secrecy

0.74

 guarded

0.71

 exits

0.71

 containment

0.68

 doors

0.67

 entrances

0.67

 Commerce

0.66

 darkness

0.65

 encl

0.65

Activations Density 0.358%

phrases related to doors and transition points

phrases related to opportunities and access

No Comments

No Known Activations