INDEX

Explanations

keywords related to technological devices and products, specifically different models

oai_token-act-pair · gpt-3.5-turbo

references to various types of models or systems

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

olulu

-0.78

azard

-0.73

vernment

-0.73

tein

-0.71

ISTORY

-0.66

 Liberties

-0.65

slaught

-0.65

 Down

-0.64

die

-0.64

orr

-0.63

POSITIVE LOGITS

model

1.24

models

1.23

 Models

1.15

 models

1.09

etter

0.86

accompan

0.83

 model

0.82

 modeling

0.78

urer

0.78

Model

0.76

Activations Density 0.013%

keywords related to technological devices and products, specifically different models

references to various types of models or systems

No Comments

No Known Activations