INDEX

Explanations

years in the 20th century

oai_token-act-pair · gpt-3.5-turbo

specific years and numerical data

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

otle

-0.81

light

-0.76

awed

-0.74

intosh

-0.74

sylvania

-0.73

ledged

-0.72

sonian

-0.72

anooga

-0.71

achu

-0.69

fall

-0.69

POSITIVE LOGITS

nd

1.93

ND

0.99

naire

0.80

 thirds

0.76

ipop

0.74

ally

0.73

nces

0.71

0.70

ppo

0.70

0.68

Activations Density 0.152%

years in the 20th century

specific years and numerical data

No Comments

No Known Activations