INDEX

Explanations

terms related to efficiency and inefficiencies

oai_token-act-pair · gpt-3.5-turbo

terms related to efficiency

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

 rally

-0.74

Ver

-0.71

mad

-0.65

Pad

-0.65

 event

-0.64

 press

-0.63

 text

-0.63

 post

-0.62

post

-0.61

ver

-0.60

POSITIVE LOGITS

iencies

5.19

iences

1.85

iency

1.79

ients

1.51

iances

1.33

ienced

1.32

ience

1.28

encies

1.27

ibilities

1.21

iration

1.18

Activations Density 0.009%

terms related to efficiency and inefficiencies

terms related to efficiency

No Comments

No Known Activations