INDEX

Explanations

words or phrases in a specific format, likely encoded messages

oai_token-act-pair · gpt-3.5-turbo

instances of dramatic elements or tension in narratives

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

 orchestr

-0.70

 distribut

-0.69

 appoint

-0.69

 redist

-0.68

 internationally

-0.67

 Burnett

-0.67

 bunker

-0.66

 geographically

-0.65

 estranged

-0.64

 allege

-0.63

POSITIVE LOGITS

Chapter

1.21

Fuck

1.06

Suddenly

1.06

ÂłÂłÂłÂł

0.95

YES

0.94

Eventually

0.92

CHAPTER

0.91

Nar

0.90

Herm

0.90

XXXX

0.90

Activations Density 0.237%

words or phrases in a specific format, likely encoded messages

instances of dramatic elements or tension in narratives

No Comments

No Known Activations