INDEX

Explanations

names and terms related to specific individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals

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

phabet

-0.74

ACTED

-0.71

Ü

-0.67

 spirited

-0.66

 suicidal

-0.63

 mileage

-0.62

 temptation

-0.62

 Scots

-0.61

 combustion

-0.61

IDES

-0.61

POSITIVE LOGITS

cz

1.31

ynski

1.19

ewski

1.06

ansky

0.99

arella

0.97

ota

0.97

oslov

0.96

ombie

0.96

owski

0.94

anski

0.94

Activations Density 0.008%

names and terms related to specific individuals

proper nouns, specifically names of individuals

No Comments

No Known Activations