INDEX

Explanations

phrases related to editorial selections or recommendations

oai_token-act-pair · gpt-3.5-turbo

references to editorial selections or features

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

 remed

-0.74

 federal

-0.72

 public

-0.66

SOC

-0.66

 residing

-0.65

 rect

-0.64

 dissemin

-0.64

 promulg

-0.62

 administrative

-0.62

 entity

-0.62

POSITIVE LOGITS

 Picks

4.49

 picks

1.83

 Pick

1.49

pick

1.39

Pick

1.36

picking

1.26

 selections

1.10

 Plays

1.10

picked

1.09

 Selection

1.03

Activations Density 0.014%

phrases related to editorial selections or recommendations

references to editorial selections or features

No Comments

No Known Activations