INDEX

Explanations

places or locations

oai_token-act-pair · gpt-3.5-turbo

references to locations and institutions

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

ãĥĨãĤ£

-0.84

 Playstation

-0.82

 mattress

-0.81

 butter

-0.81

 flats

-0.79

omnia

-0.78

Lev

-0.75

drm

-0.73

 Farrell

-0.72

 tires

-0.72

POSITIVE LOGITS

Ang

1.95

Ang

1.92

ang

1.81

ANG

1.66

Ung

1.54

ang

1.50

ANG

1.50

 Chang

1.43

 Sang

1.39

 Shang

1.38

Activations Density 0.266%

places or locations

references to locations and institutions

No Comments

No Known Activations