INDEX

Explanations

references to specific data points or information

oai_token-act-pair · gpt-3.5-turbo

parentheses and their variations in text

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

 entitle

-0.72

 hypers

-0.71

Ku

-0.71

 jump

-0.68

 electr

-0.67

 invari

-0.67

 Rocket

-0.65

 PowerShell

-0.64

 piping

-0.63

ho

-0.63

POSITIVE LOGITS

...)

1.22

âĢ¦)

1.13

emphasis

1.11

formerly

1.05

Attach

1.05

please

1.01

Introduced

1.00

?)

1.00

sic

1.00

PDF

0.93

Activations Density 0.078%

references to specific data points or information

parentheses and their variations in text

No Comments

No Known Activations