INDEX

Explanations

names or terms related to specific individuals with variations in spelling

oai_token-act-pair · gpt-3.5-turbo

proper names and entities in the 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

sburg

-0.79

 Niet

-0.72

ptr

-0.70

IVERS

-0.70

mort

-0.69

portion

-0.69

 stub

-0.69

 adolesc

-0.67

shaw

-0.67

ãģ®éŃĶ

-0.65

POSITIVE LOGITS

azi

1.07

adian

0.89

ellen

0.81

ereo

0.79

abad

0.77

ourge

0.77

ellation

0.72

olate

0.72

annis

0.70

ology

0.70

Activations Density 0.026%

names or terms related to specific individuals with variations in spelling

proper names and entities in the text

No Comments

No Known Activations