INDEX

Explanations

relationships or comparisons involving proportions

oai_token-act-pair · gpt-3.5-turbo

references to proportions and ratios in various contexts

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

 Brunswick

-0.70

 Journal

-0.67

Journal

-0.66

 Valhalla

-0.65

 Journals

-0.64

Address

-0.63

 Topics

-0.63

Rav

-0.62

 Speedway

-0.60

ISSION

-0.60

POSITIVE LOGITS

ally

1.22

ality

1.16

ate

1.11

ately

1.09

age

0.97

naire

0.91

uary

0.86

ancy

0.86

bread

0.85

bender

0.84

Activations Density 0.013%

relationships or comparisons involving proportions

references to proportions and ratios in various contexts

No Comments

No Known Activations