INDEX

Explanations

proper nouns or names

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and organizations across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 reper

-0.65

 normative

-0.62

 Citiz

-0.61

ocytes

-0.61

oplan

-0.60

drive

-0.57

íķ

-0.57

altern

-0.57

 nursery

-0.56

ocyte

-0.56

POSITIVE LOGITS

ĸļ

1.28

terness

0.84

SourceFile

0.73

vier

0.72

cliffe

0.71

zona

0.70

 Casting

0.67

taboola

0.66

bilt

0.66

VERTISEMENT

0.65

Activations Density 2.625%

proper nouns or names

references to specific individuals and organizations across various contexts

No Comments

No Known Activations