INDEX

Explanations

comments and replies in a text

oai_token-act-pair · gpt-3.5-turbo

instances of comments and responses in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

tto

-0.74

kinson

-0.72

crow

-0.69

metry

-0.68

teenth

-0.67

isSpecialOrderable

-0.66

 Takeru

-0.65

lap

-0.64

 teasing

-0.64

 celebrations

-0.63

POSITIVE LOGITS

ing

1.10

ership

1.05

Reply

1.00

erate

0.89

Comment

0.87

estamp

0.86

edIn

0.86

Reader

0.84

edin

0.84

ists

0.83

Activations Density 0.028%

comments and replies in a text

instances of comments and responses in text

No Comments

No Known Activations