INDEX

Explanations

references to theoretical concepts or academia

oai_token-act-pair · gpt-3.5-turbo

theoretically oriented ideas or concepts

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

 Annotations

-0.62

 copy

-0.61

 Foot

-0.61

 pains

-0.60

:/

-0.58

ãĥŃ

-0.58

Ms

-0.57

 clicks

-0.57

 alike

-0.57

 committed

-0.57

POSITIVE LOGITS

oret

5.04

orem

1.56

ories

1.34

ory

1.31

orie

1.27

ore

1.21

ores

1.15

oric

1.05

oresc

1.02

oren

1.01

Activations Density 0.012%

references to theoretical concepts or academia

theoretically oriented ideas or concepts

No Comments

No Known Activations