INDEX

Explanations

words related to exclamations or dialogue in a narrative context

oai_token-act-pair · gpt-3.5-turbo

capitalized letters or names, suggesting a focus on proper nouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Hancock

-0.67

 Catalyst

-0.65

 Polk

-0.65

 bilingual

-0.64

 MacArthur

-0.64

 influential

-0.63

CPS

-0.63

 Gladiator

-0.62

 Large

-0.60

 formerly

-0.60

POSITIVE LOGITS

aaaa

1.66

oooo

1.64

mmmm

1.57

OOOO

1.56

aaa

1.56

ahah

1.50

OOOOOOOO

1.49

oooooooo

1.48

mmm

1.46

AAAAAAAA

1.44

Activations Density 0.106%

words related to exclamations or dialogue in a narrative context

capitalized letters or names, suggesting a focus on proper nouns

No Comments

No Known Activations