INDEX

Explanations

words related to data and information dissemination

oai_token-act-pair · gpt-3.5-turbo

numerical data related to statistics and records

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

rawdownloadcloneembedreportprint

-0.71

needs

-0.68

comes

-0.67

reddits

-0.67

 Travels

-0.67

ESCO

-0.67

 Solitaire

-0.66

fires

-0.65

aughs

-0.64

dinand

-0.62

POSITIVE LOGITS

 indicate

1.85

 show

1.66

 suggest

1.54

 reveal

1.49

 demonstrate

1.42

 depict

1.37

 confirm

1.37

 showed

1.33

 imply

1.28

show

1.28

Activations Density 0.328%

words related to data and information dissemination

numerical data related to statistics and records

No Comments

No Known Activations