INDEX

Explanations

phrases related to actions or activities that are considered meaningful or impactful

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Norn

-0.92

 Norwich

-0.89

 McCl

-0.87

 Lars

-0.86

 Matthews

-0.84

 Maiden

-0.80

borough

-0.80

Smy

-0.80

Lind

-0.79

 Whale

-0.79

POSITIVE LOGITS

Action

1.32

action

1.30

 action

1.29

 activ

1.29

ACTION

1.26

 Action

1.26

 activity

1.24

 actions

1.21

 Actions

1.19

activity

1.18

Activations Density 0.412%

phrases related to actions or activities that are considered meaningful or impactful

No Comments

No Known Activations