INDEX

Explanations

detailed information that varies over a wide range

oai_token-act-pair · gpt-3.5-turbo

multiple subjects or topics that vary in scope or degree

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

mit

-0.78

si

-0.75

bal

-0.74

ÄŁ

-0.72

ellen

-0.69

advertisement

-0.68

scape

-0.68

zig

-0.67

rafted

-0.67

template

-0.66

POSITIVE LOGITS

 ranging

0.96

 ranges

0.94

 range

0.79

range

0.78

ranging

0.78

 Canaver

0.74

 ranged

0.71

 Luxem

0.71

 Range

0.67

ensical

0.67

Activations Density 0.014%

detailed information that varies over a wide range

multiple subjects or topics that vary in scope or degree

No Comments

No Known Activations