INDEX

Explanations

phrases indicating the return of someone from a specific location or situation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "from."

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

ounced

-0.80

retty

-0.75

acci

-0.73

ounce

-0.72

priority

-0.70

stem

-0.70

anted

-0.69

ounces

-0.68

hots

-0.68

alk

-0.67

POSITIVE LOGITS

 whence

1.05

 afar

0.94

 Defeat

0.77

 scratch

0.72

 thence

0.70

 Redemption

0.65

Rig

0.65

 captivity

0.64

 Shrine

0.64

 Reconstruction

0.64

Activations Density 0.113%

phrases indicating the return of someone from a specific location or situation

instances of the word "from."

No Comments

No Known Activations