INDEX

Explanations

words related to actions or activities

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "ACT" and variations related to actions or activities

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

nel

-0.71

kowski

-0.70

 chau

-0.69

 strips

-0.65

ley

-0.65

 Mous

-0.64

hij

-0.63

lim

-0.62

 Sloven

-0.62

 strip

-0.62

POSITIVE LOGITS

ACT

4.31

ACTION

2.32

acts

2.16

act

2.05

ACTED

1.86

ACT

1.60

acting

1.42

acted

1.39

AC

1.32

Act

1.29

Activations Density 0.007%

words related to actions or activities

occurrences of the word "ACT" and variations related to actions or activities

No Comments

No Known Activations