INDEX

Explanations

averaged values

oai_token-act-pair · gpt-3.5-turbo

phrases related to statistical averages or calculations

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

Priv

-0.68

ora

-0.63

kid

-0.63

 security

-0.62

Ãį

-0.60

dit

-0.59

ida

-0.58

 Creation

-0.57

 rescue

-0.57

 branding

-0.57

POSITIVE LOGITS

 averaged

3.55

 averaging

2.59

 averages

2.40

 totaled

1.69

 average

1.54

verages

1.45

 tallied

1.42

average

1.41

Average

1.41

 Average

1.38

Activations Density 0.017%

averaged values

phrases related to statistical averages or calculations

No Comments

No Known Activations