INDEX

Explanations

phrases related to conversations and dialogue between unidentified individuals

oai_token-act-pair · gpt-3.5-turbo

references to unidentified or uncharacterized entities, particularly around gender

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

sab

-0.68

 moons

-0.63

 fort

-0.63

sid

-0.63

san

-0.62

 manif

-0.61

 farmers

-0.61

ho

-0.59

 belt

-0.59

dru

-0.59

POSITIVE LOGITS

ABLE

1.54

ING

1.52

ITY

1.51

ERY

1.48

ISH

1.46

ED

1.46

INE

1.46

OUT

1.46

ARE

1.45

ELL

1.45

Activations Density 0.061%

phrases related to conversations and dialogue between unidentified individuals

references to unidentified or uncharacterized entities, particularly around gender

No Comments

No Known Activations