INDEX

Explanations

phrases indicating a strong opinion or warning

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the need for clarity or correction in misunderstandings

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ourses

-0.80

iliated

-0.72

ãĥŁ

-0.71

und

-0.69

onet

-0.66

affiliated

-0.66

"},"

-0.66

tumblr

-0.65

isson

-0.65

itted

-0.64

POSITIVE LOGITS

Kev

0.72

 math

0.60

!:

0.60

 Jude

0.55

 caveats

0.54

=]

0.54

 folks

0.53

mov

0.53

 Galile

0.53

,,,,

0.53

Activations Density 0.220%

phrases indicating a strong opinion or warning

phrases indicating the need for clarity or correction in misunderstandings

No Comments

No Known Activations