INDEX

Explanations

proper nouns related to individuals, places, or businesses

oai_token-act-pair · gpt-3.5-turbo

occurrences of the term "arg" followed by numeric values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

FORMATION

-0.69

 Giov

-0.69

PER

-0.67

birth

-0.65

utral

-0.64

HAEL

-0.64

 tremend

-0.64

 spaced

-0.62

 fertility

-0.62

 coli

-0.62

POSITIVE LOGITS

uably

1.43

uments

1.18

ansas

1.17

emouth

1.08

raph

1.02

roup

0.99

raphic

0.93

alos

0.91

oths

0.89

atis

0.88

Activations Density 0.017%

proper nouns related to individuals, places, or businesses

occurrences of the term "arg" followed by numeric values

No Comments

No Known Activations