INDEX

Explanations

technical terms and codes

oai_token-act-pair · gpt-3.5-turbo

references to "Shock" and related media content

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

olson

-0.90

espie

-0.74

conn

-0.68

HOU

-0.68

ezvous

-0.67

ket

-0.67

ierrez

-0.67

 fres

-0.66

 vill

-0.66

pal

-0.66

POSITIVE LOGITS

 Shock

2.31

 Become

1.76

 Contra

1.52

 Matrix

1.18

 Vector

1.17

 Transformers

1.09

 Terminator

1.06

 Infect

0.99

UCS

0.99

 Reloaded

0.97

Activations Density 0.032%

technical terms and codes

references to "Shock" and related media content

No Comments

No Known Activations