INDEX

Explanations

numeric values indicating a large amount or size

oai_token-act-pair · gpt-3.5-turbo

references to figures and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Bounty

-0.66

 overfl

-0.64

 safe

-0.63

jo

-0.63

 Affairs

-0.62

 Dawn

-0.62

 closed

-0.61

ansk

-0.60

 water

-0.60

tam

-0.59

POSITIVE LOGITS

figure

4.81

fig

3.54

figured

2.97

 Figures

1.27

FIG

1.27

Figure

1.22

igure

1.20

 figure

1.16

digit

1.11

 figures

1.04

Activations Density 0.010%

numeric values indicating a large amount or size

references to figures and statistics

No Comments

No Known Activations