INDEX

Explanations

references to a specific person or character likely named "Norman"

oai_token-act-pair · gpt-3.5-turbo

references to the character Oreman

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

 Comp

-0.71

Dat

-0.70

Lif

-0.66

SAP

-0.66

 Boost

-0.65

 Quartz

-0.64

BB

-0.63

Gam

-0.63

iP

-0.62

Sy

-0.62

POSITIVE LOGITS

orman

4.55

orm

1.38

oring

1.13

oman

1.12

ixon

1.07

oyer

1.06

olan

1.05

orney

1.02

orage

1.02

amera

1.01

Activations Density 0.002%

references to a specific person or character likely named "Norman"

references to the character Oreman

No Comments

No Known Activations