INDEX

Explanations

phrases related to enabling or causing something to happen

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate capabilities or functionalities related to various technologies or systems

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

Sum

-0.64

 rebutt

-0.63

omsday

-0.62

 wrestle

-0.62

mie

-0.62

 Taste

-0.61

 Dane

-0.59

 Jazz

-0.58

Nab

-0.58

 Schne

-0.58

POSITIVE LOGITS

 enabled

3.57

enabled

2.50

 Enabled

2.00

 enable

1.95

Enabled

1.86

 enables

1.66

 disabled

1.62

 enabling

1.52

enable

1.42

 Enable

1.40

Activations Density 0.012%

phrases related to enabling or causing something to happen

phrases that indicate capabilities or functionalities related to various technologies or systems

No Comments

No Known Activations