INDEX

Explanations

phrases related to separating or destroying something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "apart" and related contexts of separation or disconnection

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

DAY

-0.69

occ

-0.69

activated

-0.68

car

-0.60

authorized

-0.60

externalActionCode

-0.60

ãĤ§

-0.60

morning

-0.60

rection

-0.58

 amen

-0.58

POSITIVE LOGITS

heid

1.11

ĸļ

0.86

ively

0.81

iveness

0.73

igious

0.70

ellig

0.69

heartedly

0.69

 sexes

0.68

chool

0.67

inguished

0.66

Activations Density 0.012%

phrases related to separating or destroying something

occurrences of the word "apart" and related contexts of separation or disconnection

No Comments

No Known Activations