INDEX

Explanations

names or terms likely related to legal or formal documents

oai_token-act-pair · gpt-3.5-turbo

references to names, particularly in a context related to discussions or analyses

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 prism

-0.68

ario

-0.65

rod

-0.64

pole

-0.64

EGA

-0.63

azo

-0.63

pport

-0.62

 solitary

-0.61

 Barg

-0.60

Pen

-0.58

POSITIVE LOGITS

names

3.30

 namespace

1.34

Names

1.27

 Names

0.92

paces

0.88

maps

0.86

template

0.85

pages

0.82

chat

0.82

stocks

0.81

Activations Density 0.029%

names or terms likely related to legal or formal documents

references to names, particularly in a context related to discussions or analyses

No Comments

No Known Activations