INDEX

Explanations

phrases indicating continuation or elaboration on a topic

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of "Article Continued Below" phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

reb

-0.72

RAW

-0.68

 Reborn

-0.64

romy

-0.63

rament

-0.62

 reneg

-0.59

formation

-0.59

 Noir

-0.58

icion

-0.56

 undermin

-0.56

POSITIVE LOGITS

 Copy

0.67

↵

0.67

 Crossref

0.66

idth

0.63

 Continue

0.63

ached

0.62

 Close

0.61

<|endoftext|>

0.61

 Below

0.61

etitive

0.61

Activations Density 0.013%

phrases indicating continuation or elaboration on a topic

repeated mentions of "Article Continued Below" phrases

No Comments

No Known Activations