INDEX

Explanations

reasons or explanations provided in text

oai_token-act-pair · gpt-3.5-turbo

phrases that describe reasons or justifications

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

OLOGY

-0.81

chn

-0.78

nets

-0.76

feld

-0.73

astics

-0.73

anges

-0.73

chin

-0.71

jab

-0.70

forts

-0.69

ros

-0.69

POSITIVE LOGITS

why

1.03

 inaction

0.99

 preferring

0.98

 existence

0.86

 excluding

0.83

 optimism

0.83

 hesitation

0.83

 dismissing

0.82

 omission

0.82

 choosing

0.81

Activations Density 0.100%

reasons or explanations provided in text

phrases that describe reasons or justifications

No Comments

No Known Activations