INDEX

Explanations

names of people or places

oai_token-act-pair · gpt-3.5-turbo

specific names or terms associated with a particular individual or character

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

 convertible

-0.72

 mailbox

-0.69

 conversions

-0.66

 clocks

-0.65

 Exodus

-0.63

 frying

-0.62

 latt

-0.61

 exempt

-0.60

fo

-0.60

 Convert

-0.60

POSITIVE LOGITS

nar

4.79

nir

1.46

Nar

1.39

nen

1.39

nas

1.34

ni

1.31

nor

1.23

nard

1.20

nik

1.19

nian

1.19

Activations Density 0.013%

names of people or places

specific names or terms associated with a particular individual or character

No Comments

No Known Activations