INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to a specific individual or name

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

STD

-0.74

vt

-0.72

UAL

-0.69

EMA

-0.68

©

-0.67

diagn

-0.67

 Turing

-0.65

 preference

-0.64

 bunker

-0.64

 vets

-0.63

POSITIVE LOGITS

Ol

3.77

Ol

2.45

ol

1.86

OL

1.45

Ole

1.42

 Olson

1.38

Osw

1.29

 Oliver

1.28

Oss

1.25

 Olsen

1.24

Activations Density 0.013%

proper nouns

references to a specific individual or name

No Comments

No Known Activations