INDEX

Explanations

words related to anticipation or predictions

oai_token-act-pair · gpt-3.5-turbo

the term "Ant" and its various forms or related mentions

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

 repatri

-0.70

 pony

-0.68

gre

-0.66

 Powell

-0.66

 numbered

-0.65

 Kyoto

-0.65

 format

-0.65

 repayment

-0.64

 backlog

-0.64

CGI

-0.63

POSITIVE LOGITS

Ant

3.84

Ant

1.97

Anth

1.44

 Anton

1.42

Anti

1.35

Aust

1.28

ant

1.28

ANT

1.27

Benz

1.27

Anthony

1.26

Activations Density 0.012%

words related to anticipation or predictions

the term "Ant" and its various forms or related mentions

No Comments

No Known Activations