INDEX

Explanations

phrases related to specific entities or organizations

oai_token-act-pair · gpt-3.5-turbo

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

 tremend

-0.81

quo

-0.78

 clock

-0.77

 heels

-0.76

 downhill

-0.74

 depend

-0.74

 slowing

-0.74

pse

-0.73

cru

-0.73

 ultr

-0.73

POSITIVE LOGITS

 Cities

1.13

 Feet

1.12

 Tide

1.10

 Moves

1.09

Mind

1.08

 Lines

1.04

 Mouth

1.03

 Investigations

1.03

 Tracks

1.01

 Horizons

1.01

Activations Density 0.146%

phrases related to specific entities or organizations

No Comments

No Known Activations