INDEX

Explanations

concepts that are correspondingly related or matching

oai_token-act-pair · gpt-3.5-turbo

terms related to correspondence and accompanying information

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

icz

-0.78

cher

-0.77

joice

-0.76

spe

-0.73

spect

-0.72

hers

-0.72

bers

-0.70

Bay

-0.70

ctors

-0.68

uve

-0.67

POSITIVE LOGITS

ively

0.72

 amounts

0.72

 periods

0.71

alties

0.68

 Pengu

0.68

 acronym

0.67

 colours

0.67

 figures

0.66

 corresponding

0.66

 colors

0.65

Activations Density 0.010%

concepts that are correspondingly related or matching

terms related to correspondence and accompanying information

No Comments

No Known Activations